In: Food Chemistry Research Developments Editor: Konstantinos N. Papadopoulos, pp.

ISBN 978-1-60456-262-0 © 2008 Nova Science Publishers, Inc.

Chapter 5

Protein and Lipid Recovery from Food Processing By-Products Using Isoelectric Solubilization/Precipitation Jacek Jaczynski West Virginia University Animal and Nutritional Sciences Morgantown, WV 26506

Abstract The isoelectric behavior of food proteins has been well characterized in the food science literature. The isoelectric point (pI) of a protein is a pH at which the protein maintains a zero net electrostatic charge. Therefore, the protein at its pI exhibits the least solubility; however, as the pH is changed, the protein-water electrostatic interactions increase; and consequently, the protein becomes water soluble. This fundamental phenomenon has been used in the food industry for a long time as for example in the cheese-making and tofu manufacture. However, more recently several food science laboratories have initiated an extensive research on the application of the pI to recover functional muscle proteins, particularly fish myofibrillar proteins. The pI of fish muscle proteins is typically at pH 5.5 and the protein solubility greatly increases above pH 10.0 and below pH 3.0. It has been reported by numerous popular media that fish stocks are declining and several commercial fisheries are currently over-exploited and will collapse by the mid century. In general, processing of raw materials for human food products inevitably entails generation of some quantities of processing by-products. However, fish processing typically results in very high amounts of by-products (heads, frames, viscera, and etc.) that are land-filled or ground-and-discarded. When fish are mechanically processed for fillets, the recovery yields are typically 30-40% of fillets and the byproducts account for 60-70%. The 60-70% of by-products contains highly nutritious fish muscle proteins and fish oil rich in heart-friendly omega-3 fatty acids. The proteins and oil could be recovered and used subsequently in the development of human food products and dietary supplements.

2

Jacek Jaczynski Unfortunately, due to the lack of a proper, commercially available technology to recover proteins and lipids from the fish processing by-products, this tremendous natural resource along with some alternative aquatic species such as Antarctic krill are not utilized for human consumption. However, the isoelectric solubilization/precipitation allows recovery of fish proteins and oil from otherwise useless fish processing byproducts at yields approximately 90%. The recovered proteins retain their functional properties such as gelation and waterholding-capacity (WHC); and therefore, can be a useful, main ingredient in the development of human food products. The amino acid profile of the recovered proteins shows that all of the essential amino acids are retained in the recovered proteins at adequate levels, demonstrating high nutritional quality. The omega-3 fatty acids in the recovered oil do not undergo degradation during isoelecteic processing; and therefore, retain their full potential for human health benefits. The commercialization of the isoelectric solubilization/precipitation has been initiated and likely the multi-fold benefits, including both economic and environmental are forthcoming. This chapter describes basic chemical principles of the isoelectric behavior of food proteins as it pertains to their structure and interaction with water. In addition, this chapter covers the most recent developments regarding application of isoelectric point to recover nutritious muscle proteins and oil from fish processing byproducts.

Introduction Since prehistoric times fish have been on the menu for hominids and subsequently for Homo sapiens (Stewart, 1994). A fish salad dated 1300 B.C. made from spiced and marinated carp in China was the first recipe ever recorded by men (Zugarramurdi et al., 1995). Simple processing methods such as smoking, fermentation, salting, and drying were the basis for the integrated industry in the Roman Empire approximately 100 B.C. (McCann, 1988). China has a long tradition in fish aquaculture. As early as 2000 B.C. fish were raised in China for food (Kreuzer, 1974). This long-standing tradition and knowledge gained over the years have probably significantly contributed to the fact that currently the world largest aquaculture production is in China (Anonymous, 2004). Processing of raw foods into food products inevitably entails generation of some types of by-products and processing aquatic foods (i.e., seafood) is no exception in this regards. In fact, utilization of the aquatic foods by-products is more important for economical viability of this sector than most other sectors (Gildberg, 2002). In traditional non-industrialized fisheries where most of the labor was provided by skilled and experienced personnel often passed down by generations, most of the fish were almost completely utilized for human consumption, animal feed, or plant fertilizer. The economy-driven industrialization of fisheries has brought tremendous development, but at the same time, the amounts by-products generated during processing have increased accordingly (Gildberg, 2002). Typical examples are commercial shrimp trawling, krill processing, or fish filleting. It is not unusual in shrimp trawling that 90% of the total catch volume is by-catch, which most of the time is discarded (Raa and Gildberg, 1982). The recovery yield of meat from Antarctic krill (Euphausia superba) in commercial processing is extremely low; between 10-15% by weight of whole krill (Suzuki, 1981). When fish are mechanically processed for fillets, the recovery yields are typically 30-40% of fillets

Protein and Lipid Recovery from Food Processing By-Products…

3

and the by-products account for 60-70% by weight of whole fish (Gildberg, 2002, Chen and Jaczynski, 2007a). While it is not an uncommon to “grind-and-dump” the 60-70% of the byproducts, the “grind-and-dump” procedure should be considered an irresponsible utilization of our natural resources. There is some misunderstanding as to how the aquatic foods processing “by-products” are defined. Three terms are frequently and interchangeably used in the fish processing industry as well as scientific literature to describe the same materials during processing of aquatic foods. They are: (1) “offal”, (2) “waste”, and (3) “by-products”. The first two terms imply that those materials cannot be used for any application and should be disposed of; and therefore, are often misleading and trigger a negative connotation. While the third term suggests that there may be some valuable components to be recovered if properly treated; and therefore, it is a positive term and as such should be used. Currently, the most common definition of the “by-products” is all of the edible or inedible materials left over following processing of the main product. A typical example is fish filleting to recover boneless and skinless marketable fillets. The fillets would be considered the main product and the frames, heads, and guts would be typical “byproducts”. As mentioned above, the amount of filleting by-products is about two times higher than that of the fillets. It would be misleading to name these by-products as “waste” since at the time of filleting the quality of fish meat left on the frame (i.e., fish bones) and in the heads is not compromised (Strom and Eggum, 1981). If proper meat recovery technology is successfully applied, the recovered meat can result in added revenue for a processor as well as reduce environmental pollution associated with disposal of the processing by-products. The by-products can be used as a basis to derive three major groups of products. They are: (1) plant fertilizers, (2) livestock feeds, and (3) human value-added foods and specialty foods. In general, conversion of the by-products to fertilizers (1) results in the least value addition to the by-products, while the value addition is the highest when human foods or specialty foods (3) are developed from the processing by-products. It has been estimated that the value addition of the human food products developed from the processing by-products will increase fivefold within the next decade (Gildberg, 2002).

Current Fisheries and Processing By-Products While the global fish capture and aquaculture production are available from the Food and Agriculture Organization (FAO) Fisheries Department (Anonymous, 2004), the amounts of byproducts can only be estimated. Although the capture fisheries have remained fairly stable for more than the past twenty years and it has been forecast that the capture fisheries are not likely to increase in the future, the world aquaculture sector has increased its production tremendously for the same period and currently contributes approximately 50% of the capture fisheries (Vannuccini, 2004) (Table 1). The FAO also predicts that the future demand for aquatic foods will have to be met by increased aquaculture production. In 2002, about 76% of the world fisheries production was used for direct human consumption and the remaining 24% was reduced mainly to fishmeal and oil (i.e., reduction fisheries).

4

Jacek Jaczynski Table 1. Statistics of world fisheries production and utilization in million tons (Source: Adapted from Anonymous (2004)) 1998

1999

2000

2001

2002

2003

Fish supply for human consumption per capita

15.8

15.9

15.9

16.2

16.2

16.3

Total human consumption

93.6

95.4

96.8

99.5

100.7

103.0

Other than food utilization

24.6

31.8

34.2

31.1

32.2

29.2

Total fisheries

118.2

127.2

131.0

130.7

133.0

132.2

Total capture fisheries

87.7

93.8

95.5

92.9

93.2

90.3

Total aquaculture production

30.6

33.4

35.5

37.8

39.8

41.9

However, while on weight basis the reduction fisheries account for the quarter of the total fish utilization, they only contribute 3.8% to the fisheries total economic value (Anonymous, 2004). Therefore, it is clear that conversion of the raw aquatic foods for human food products instead of animal feeds is more economically beneficial on global scale. This is why food products developed from the processing by-products should be considered typical value-added foods. Utilization of fish processing by-products by the rendering industry yields fishmeal and oil essentially without any subsequent by-products. However, about 100 million tons is directly consumed by human population (Table 1). Commercial filleting of fish such as cod, salmon, trout, tilapia, seabream, pollack, etc. typically yields about 60-70% of by-products and 30-40% fillets of whole fish weight. In general, fish meat and oil left on the by-products following filleting may range widely, but typically accounts for 20-30 and 5-15% of whole fish weight, respectively (Chen and Jaczynski, 2007a; Chen et al., 2007b). Fish oil is highly polyunsaturated (omega-3 fatty acids (ω-3 FA)), and therefore, very susceptible to lipid oxidation, resulting in rancidity development (Chen et al., 2006; Chen et al., 2007a). If fish oil could be efficiently isolated from the by-products and used for human consumption in a form of for example encapsulated dietary supplements, this would likely result in an additional stream of revenue for a processor. Fish by-catch and discards fall into the by-product category. The by-catch is particularly high in shrimp reaching 90% of the total catch. However, on average, the by-catch and discards account for 30% of the total world capture fisheries, which translates into approximately 30 million tons of available resource, yet not utilized (Zugarramurdi et al., 1995). Another aquatic resource that is not commercially utilized at a full scale for human consumption is krill (Figure 1).

Protein and Lipid Recovery from Food Processing By-Products…

5

Figure 1. The sustainable biomass of Krill that could be used for human consumption on annual basis has been estimated to be larger than the world annual consumption of all other aquatic species.

The only commercial utilization of krill is the reduction fisheries to manufacture fish feed. However, if proper technology is developed to efficiently convert this resource to human food, a significant value addition could result. At the same time, krill could contribute significantly to fulfill human nutritional needs for aquatic proteins; and therefore, alleviate some environmental stress associated with over-fishing and depletion of some marine species. This tremendous resource has been estimated at 400 – 1550 million tons with sustainable annual krill harvest being around 70 – 200 million tons a year (Suzuki and Shibata, 1990). Therefore, krill biomass that could be available for human food is comparable to the biomass of all of the other aquatic species currently harvested by human. The krill biomass is probably the biggest of any multicellular animal species on the planet (Nicol and Endo, 1999). Krill are not full utilized for human consumption due to the lack of efficient meat recovery technology. Krill are small crustaceans that resemble shrimp. However, krill unlike shrimp, have extremely active proteolytic enzymes. During harvest these enzymes are released and literally liquefy krill meat at rapid rates (Kolakowski and Lachowicz, 1982).

Opportunity to Develop Protein and Lipid Recovery Technology The biomass of aquatic by-products and underutilized species such as for example krill is staggering. At the same time, over-fishing, stock depletion, and other environmental issued associated with processing of aquatic foods have become increasingly more emphasized in popular media. Unlike aquatic foods from our natural resources, the world population is increasing. Therefore, it may be increasingly more difficult to meet human nutritional needs for proteins and lipids from aquatic resources in the future. Although aquaculture may partly alleviate these issues by providing more raw materials for processing (input), new and more efficient recovery technologies will have to be developed

6

Jacek Jaczynski

and implemented in order to increase recovery yields of human food products (output) and to reduce the amounts of by-products (output minus input). A challenge will also be to find proper applications for recovered materials, so that the processing would be economically sustainable. In order to develop efficient recovery technologies it is necessary to understand some fundamental properties of the raw materials (input). By understanding these properties one may be able to manipulate behavior of the bio-molecules and in turn, increase the recovery yields (output). Meat is the edible portion of aquatic animals that is derived from a muscle tissue. Meat is a mixture of mainly water, proteins, and intra-muscular lipids. Therefore, basic understanding of these three components will be necessary to be able to develop new recovery technologies or to maximize current recovery yields.

Water Properties in Foods Water is a prevalent constituent of meat and depending on the species, the water content may range very widely, reaching values up to 90%. In a final food product derived from aquatic species, not only does the water control the weight of the product; and therefore, the revenue, but also water content affects how succulent and juicy (sensory attributes) the aquatic foods are perceived by customers. This is why it is important to be able to control water retention in aquatic foods. A water molecule is relatively simple and small compared to other bio-molecules and it is composed of hydrogen and oxygen. Since water does not contain carbon atom, it is considered inorganic. From a human nutrition standpoint, water unlike lipids, proteins, or carbohydrates does not contribute calories in human body. Therefore, from a processor prospective, the higher the water content, the “lighter” (i.e., less calories) the product. Water is often added during formulation stages to “dilute” the energy-yielding compounds (i.e., lipids, proteins, and carbohydrates). However, when water is added to the product, water-binding compounds may need to be used to prevent “drip” losses. Meat (muscle) proteins have very good water holding capacity (WHC) and if raw aquatic foods have not been abused, particularly thermally, then they can retain water very well. For example, krill muscle proteins have very good WHC, however, they are very sensitive to temperature abuse (Suzuki, 1981). In meat systems, water provides a reaction medium in which all other compounds such as proteins and lipids may be dissolved or suspended. Therefore, it is critical to be familiar with how water can interact (i.e., dissolve) with these compounds in meats derived from aquatic animals. A molecule of water is commonly called a dipole due to its slight electro-negative and electro-positive charges of oxygen and hydrogen atoms, respectively (Figure 2). This occurs because the larger oxygen atom is powerful enough to attract the electrons from the two hydrogens; and therefore, causing a slight shift of electronegativity towards oxygen, which consequently makes hydrogen a little more positively charged. This means that water molecules have electrostatic charges on their surface. These charges allow interaction between water dipoles and other charges molecules such as proteins (Figure 2).

Protein and Lipid Recovery from Food Processing By-Products…

7

Figure 2. Protein solubility is due to electrostatic charges of water dipoles. Due to the charge shift water dipoles are capable of creating weak bonds between each other or other charged molecules such as for example proteins in meat systems.

This property is very important in food technology, because many major compounds in meats such as proteins may also be charged; and therefore, have an ability to interact with water dipoles via these interactions. This is why a prerequisite for a meat component to be water soluble is that the component must be electrostatically charged in order to bond with the water dipole. This is the basis for water solubility of many food compounds.

Basic Properties of Fish Proteins and Lipids Fish Proteins There are two basic types of muscles present in raw aquatic foods (1) striated muscle is a major component of meat derived from fish and terrestrial animals, and (2) smooth muscle is characteristic of meat derived from molluscans. Fish muscle can be divided further into white and dark muscle. The dark muscle lies alongside fish body and under the skin. The striations of fish muscle are very distinct. The muscle fibers are mainly composed of parallel myofibrils with the space in between filled by sarcoplasm (i.e., sarcoplasmic proteins). The myofibrils are made of two major myofilaments that vary by thickness. The thick and thin filaments are called myosin and actin, respectively. They are main meat proteins and myosin is largely responsible for meat functionalities. The bundles of myofibrils are held together by connective tissue (i.e., stroma proteins). However, fish meat unlike meat derived from terrestrial animals has in general much less of the connective tissue (fish – 3-5%, beef – 16-28%) (Suzuki, 1981). Ionic strength (IS) is an important factor in water solubility of the meat proteins. The IS represents the status of electrolytes (i.e., ionized salts) present in a system. Generally, the higher the IS, the more salt is present is a meat system. Provided the IS is not modified, the two major myofibrillar proteins, myosin and actin are water insoluble. Myosin and actin have been

8

Jacek Jaczynski

shown to be water soluble at extremely low IS or above 0.6. While the sarcoplasmic proteins are quite water soluble, the stroma proteins are completely water insoluble. The water solubility of sarcoplasmic proteins decreases with the increase of IS (Lanier et al., 2005; Suzuki, 1981). Proteins, including fish muscle proteins are made up of amino acids (AA). There are about twenty AA. While all of the AA are indispensable for our normal growth and sustaining our metabolic processes, nine of them cannot be synthesized by adult human. This is why these nine AA are named “essential” (EAA). It also means that the only way we can acquire these nine EAA is adequate diet. The biological value (BV) of a protein measures its efficiency in supporting the body’s needs. Egg proteins are regarded as a reference source and have BV of 100, meaning that 100% of the nitrogen is absorbed and retained. Milk, beef, fish, corn, and rice proteins have BV of 93, 75, 75, 72, and 59, respectively (Whitney and Rolfes, 2005; Murano, 2003). The BV of Krill proteins has been reported as slightly higher than the BV of milk proteins (Suzuki and Shibata, 1990). As in their name, all of the AA contain amino and acid (carboxyl) groups that are bonded to a central carbon atom (Figure 3). These amino and acid groups, often referred to as functional groups in food science, create a strong covalent bond. This specific bond is called a peptide bond. Several AA bonded together by the peptide bond make a protein. Unlike the bond that holds the water dipoles together or results in protein-water interactions, the peptide bond is much stronger. These amino and acid groups can also be electrostatically charged in solutions; and therefore, participate in protein-water interactions via weak hydrogen bonds (i.e., water solubility) (Figure 3). The bonding energy of a hydrogen bond is low; however, when present in high numbers, these bonds efficiently stabilize the complex three dimensional structures of food proteins such as meat proteins. The amino and acid groups bonded to the central carbon atom are a core of the protein molecule; however, what gives the uniqueness and in turn results in the twenty different AA available is the side chain (group) (Figure 3). The side chain is unique for each AA, and in fact results in vastly diverse characteristics of different food proteins. Depending of the side chain attached to the core (i.e., central carbon atom), the protein molecule can have different properties. While hydrophobic side chains result in limited water solubility of the protein, the polar (charged) side chains may result in considerable proteinwater interaction via the hydrogen bond; thereby allowing water solubility. However, the weak hydrophobic interaction may result in protein-protein interactions and their aggregation, followed by precipitation (Figure 3). The protein-water interaction is essential for protein solubility as well as water holding capacity (WHC). The side chain of one AA, cysteine contains a sulfhydryl group (–SH). When the –SH groups are oxidized (i.e., during heat treatment), a covalent bond called di-sulfide bond (–SS–) is created. The strong –SS– bond in combination with the weak hydrophobic interactions between the hydrophobic AA are essential in gelation of fish proteins and are responsible for proper texture development of final food products such as for example value-added foods. The side chains can also be chemically modified, resulting in modified proteins that exhibit different functionalities than their parent molecules. Although these chemical modifications may make proteins unsuitable for human consumption, the unconventional non-food applications may be developed for the proteins recovered from the food processing by-products.

Protein and Lipid Recovery from Food Processing By-Products…

9

Figure 3. Proteins, including meat proteins of aquatic foods, are composed of amino acids (AA). The AA are made up of amino (+NH3) and acid (COO–) groups bonded to a central carbon atom and side chain (group) (R). The side chain is unique for AA and may allow proteins to create protein-protein hydrophobic interactions and protein-water hydrogen bonds.

Depending on conditions to which the fish proteins are subjected, the protein side chains can assume different electrostatic charges (Figure 4). This means that the protein solubility can be “turned” on or off by providing conditions which either favor or disfavor protein solubility, respectively. By adding acid to a solution, the acid dissociates yielding positive hydrogen ions, H+ (i.e., hydronium ion, H3O+). These positive charges interact with the negative charges on the protein; thereby, neutralizing them and making the protein more positively charged. By adding a base to a solution, the protein eventually assumes more negative charge on its surface. The general mechanism for base addition is similar to that with the acid; however, the final protein charge will be negative due to the base instead of positive that is imposed by an acid (Figure 4).

10

Jacek Jaczynski

Figure 4. Isoelectric point (pI) of a protein is a pH at which the net electrostatic charge of the protein equals zero (neutral). If a protein is at its pI (left side of the figure), the water solubility of the protein (protein-water interactions) is at its minimum; while protein-protein interactions are the maximum, resulting in protein-protein interactions via weak hydrophobic bonds. Therefore, the protein precipitates. If a protein is shifted from its pI (acidic or basic conditions) (right side of the figure), the protein-water interactions become prevalent, resulting in increased protein solubility in water.

As a protein assumes more overall positive or negative charge, it gradually starts electrostatic interaction with water dipoles (i.e., protein-water intearctions). As protein-water interactions increase, the protein-protein hyrdrophobic interactions decrease. Therefore, as the protein molecules become more polar (charged), their water solubility increases and they become water soluble. However, it is possible to adjust pH of a protein solution at which the number of negative charges on a protein’s surface is equal to the number of positive charges; and therefore, the protein molecule assumes an overall neutral electrostatic charge. The pH at which the overall electrostatic charge of a protein is equal zero is called isoelectric point (pI) (Figure 4). The pI is very specific for different proteins. It is very important from a food technologist/scientist standpoint to know the pI of proteins, because as the charges on protein’s surface diminish, so do the protein-water interactions; and hence, water solubility. However, the hydrophobicity driven protein-protein interactions are favored at the pI; and therefore, proteins precipitate at their pI. This basic chemical pH-induced protein behavior allows modification of protein solubility/precipitation by proper pH adjustment. As proteins are a major constituent (besides water) of fish meat, this property can be used to recover fish meat proteins from aquatic foods processing by-products. While fish meat proteins are in a soluble form (protein-water interactions are favored), the impurities (bone, skin, scale, etc.) can be removed from the solution, followed by protein precipitation and recovery at their pI (proteinprotein interactions are favored). Milk protein, casein or soy proteins are recovered from their

Protein and Lipid Recovery from Food Processing By-Products…

11

raw sources by the application of this principle and further used in various food and non-food applications. A good example is cheese where casein is precipitated from milk by adjusting pH to 4.6, resulting in a cheese curd (i.e., precipitated casein) and subsequent draining of whey. Similar to other food proteins, fish proteins may loose their functionalities if protein denaturation occurs. Denaturation is a complex physico-chemical phenomenon that results in spatial rearrangement of the native, three dimensional protein structures. This rearrangement is caused primarily by the disruption of the hydrogen bonds that stabilize the native protein structure. If the native structure is significantly modified (i.e., denatured), the protein will loose for example its solubility and WHC; and consequently will make a weaker gel (i.e., a product with poor texture and water drip). Denaturation may be reversible or irreversible depending on the level and intensity of a denaturing agent. Typical denaturation occurs if proteins are thermally abused, for example when freshly caught fish are not handled properly, the products derived from these fish will have poor quality. Another example is prolonged frozen storage or inadequate freezing temperatures of fish and fish products. In general, proteins derived from cold water fishes are more sensitive to the temperature abuse than their counterparts from warm waters. Ionizing radiation and pH are other examples of denaturing agents. These functional properties are very important for food technologists and are the biochemical principles for the protein recovery system described in the subsequent sections. Therefore, their thorough understanding is critical.

Fish Lipids The fat content in fish muscle is highly variable. It depends on species, age, spawning season, fish diet, and the body part. While the protein content in fish muscle is relatively constant, the fat content is typically inversely correlated to the moisture content. It means that the more fat in the fish meat, the less water in the muscle. Due to fish metabolism the dark fish muscle contains more fish oil than the white muscle. Not only does the dark muscle have more oxygen supplied by the blood to deteriorate the oxidation-sensitive fish oils, but the dark muscle also has more pro-oxidative hemoglobin and mitochondria (organelle that oxidizes energy-yielding substrates). Therefore, the dark muscle is more prone to lipid oxidation and development of off-odors (i.e., rancidity) (Hultin et al., 2005). For these reasons, the white flesh fish are preferred by customers; and hence, processed by the industry. However, it would be economically beneficial to develop a technology capable of fish oil removal from the dark fish meat; and thereby, adding some value to the lower grade products. The intramuscular lipids (including fish oils) of importance to food technologists are mainly composed of glycerol backbone covalently bonded via ester bond to three fatty acids (FA) chains of various lengths and saturation levels (Figure 5). Because of the number of the FA attached to the glycerol, the intramuscular lipids are often referred to as triglycerides.

12

Jacek Jaczynski

Figure 5. Major intramuscular (meat) lipids are made of glycerol backbone to which three fatty acids (FA) are attached via strong covalent bond; and hence the name triglycerides.

The functional properties of fish oils depend on the composition of FA attached to glycerol. Overall, lipids are hydrophobic and lighter (i.e., lower specific density) than water; and therefore, when fish oils are in water, they tend to join together (i.e., coalesce) and float to the top if enough time is allowed or sufficiently high g force is applied. Unlike fish muscle proteins, fish oil (i.e., triglycerides) does not have charges present on their surface and cannot bond with water. For this reason lipids are frequently called apolar compounds. Due to hydrophobic characteristics of lipids, they can interact with hydrophobic side chains (R) of AA in fish muscle proteins; and therefore, create weak lipid-protein hydrophobic bonds. Another type of lipids that are probably more important to fish technologists than triglycerides are phospholipids. Phospolipids are an integral component of cell membranes. Phospholipids are also made of the glycerol backbone; however, instead of three FA only two are covalently attached via ester bond to the glycerol backbone. The third place on the glycerol backbone is filled with a positively charged phosphate group linked with some other charged moiety (most of the time choline) (Figure 6). Due to the charge, phospholipids have dual properties. The compounds with dual lipo- and hydro-philic properties are often called amphiphilic compounds. Phospholipids can create hydrophobic bonds with other apolar substances (for example fish oil) and at the same time they can interact with water and/or

Protein and Lipid Recovery from Food Processing By-Products…

13

charged proteins (for example fish muscle proteins) (Figure 6). Therefore, phospholipids are typical examples of emulsifiers. When they are present in a solution during processing of aquatic foods (i.e., minced fish meat), separation of fish oil from water and proteins is more difficult.

Figure 6. Similarly to triglycerides, phospholipids can form hydrophobic bonds with apolar substances. However, phospholipids can also interact with other charged food molecules such as charged proteins and water dipoles due to the charge present on the phospholipid molecule. Therefore, fish membrane lipids unlike fish oil (i.e., phospholipids) are water soluble.

The membrane phospholipids tend to have higher level of unsaturation as well as greater surface area than muscle triglycerides. In addition, membrane phospholipids are also often in a close association with pro-oxidative processes such as those in mitochondria. Therefore, membrane phospholipids are more susceptible to oxidation than intramuscular triglycerides. Although the content of phospholipids is lower than the content of triglycerides in fish muscle, the phospholipids due to their properties contribute more to the rancidity development (Hultin et al., 2005). The problem is exacerbated because the phospolipids are difficult to separate from minced fish due to their amphiphilic characteristics (lipo- and hydro-philic). Fish oils compared to the fats of terrestrial origins have higher level of unsaturation. Although the polyunsaturated fatty acids (PUFA) have been correlated with improved cardiovascular health of human, the PUFA are highly susceptible to lipid oxidation that leads to fish rancidity (Chen et al., 2006; Chen et al., 2007a). Most of the human health benefits have been ascribed to the eicosapentaenoic (EPA, 20:5ω-3) and docosahexaenoic (DHA, 22:6ω-3), which has resulted in numerous products supplemented with these FA. While the DHA is commonly used to fortify infant formula in the U.S., the fish and fish products are excellent natural sources of this “heart-friendly” FA. Both the DHA and EPA are examples of omega-3 (ω-3) FA. The ω-3 FA are considered “essential” (EFA) from human nutrition stand point,

14

Jacek Jaczynski

because similarly to the essential amino acids, the EFA cannot be synthesized by human body. Therefore, the EFA must be provided in the diet for proper development and functioning. The ω-3 nomenclature refers to the third carbon atom in the FA chain where the first double bond (C=C) occurs counting from the methyl end (CH3) of the FA (Figure 5). In general, the fish oils are highly valued by customers, as demonstrated by their frequent inclusion in dietary supplements and functional foods. However, at the same time, if fish are not properly processed and stored, the oxidation of fish lipids is rapid leading to the onset of rancidity. Many customers find fish rancidity offensive. Since lipid oxidation is a chemical reaction, a thorough understanding of the oxidation events that fish oils undergo allows a development of proper strategies to prevent this detrimental reaction. Common strategies include vacuum packaging (lack of oxygen), antioxidants (example: tocopherols – vitamin E), oxygen scavengers, freezing, and etc. (Chen et al., 2007a). Fish lipids are commonly referred to as fish oils due to their low melting point. The melting point mainly depends on the composition of fatty acids that are attached to the glycerol. Fish oils due to high concentration of the long chain (LC) PUFA tend to have lower melting point than their counterparts derived from terrestrial animals such as lard or suet. This is why fish oil is in a liquid state at room temperature. However, fish are often processed at cold temperatures; and therefore, fish oil becomes more viscous. In order to facilitate separation of fish oil from minced fish meat, a processor may increase temperature slightly – just enough to make the oil less viscous; and therefore per Stoke’s law, facilitate separation under g force in a centrifuge. This property is used when raw milk is slightly heated before cream is separated during the manufacture of skim milk. However, it needs to be pointed out that the FA in milk are mainly saturated; and therefore, less susceptible to oxidation. As phospholipids are active components of semi-permeable cell membranes that require high level of fluidity for their proper functioning, low melting point of fish phospholipids correlates well with the relatively colder fish habitat when compared to their terrestrial counterparts.

Recovery of Functional Proteins and Lipids from Aquatic Foods Processing By-Products by Isoelectric Solubilization/Precipitation at Pilot Scale Overview The principle of the isoelectric point (pI) has long been used in cheese making and manufacture of soy protein isolates. Major milk protein, casein precipitates at its pI following acidification of milk to the pH = 4.6 by the action of rennet and/or lactic acid bacteria (LAB). This acidification results in a formation of curd from casein. The pI of fish muscle proteins is 5.5. Therefore, fish muscle proteins precipitate at the pH = 5.5 and become gradually watersoluble as the pH is changed to either acidic or basic. The ionic strength (IS) is an important factor in water solubility of the fish muscle proteins. The isoelectric solubilization/precipitation of fish muscle proteins with concurrent separation of fish oil was proposed by food scientists from the University of Massachusetts (Hultin and Kelleher, 1999, 2000, 2001, 2002).

Protein and Lipid Recovery from Food Processing By-Products…

15

Following pioneering developments by Hultin and Kelleher along with earlier works by Meinke et al. (1972, 1973), several food science laboratories in the U.S. and abroad started active research in this field. In general, there are five steps in the protein and lipid recovery based on isoelectric solubilization/precipitation (Figure 7). The first step is to homogenize (i.e., grind) the byproducts with water at 1:6 (wt/wt) ratio in order to provide reaction medium and increase surface area for the subsequent protein solubilization reaction.

Figure 7. A general flow chart of the protein and lipid recovery system using isoelectric solubilization / precipitation. Materials in the boxes are the recovered fractions that can be further processed for food and non-food applications.

In the second step, the fish muscle proteins are solubilized at either acidic or basic conditions. As the pH moves further away from the pI, the fish muscle proteins assume more uniform either negative or positive surface electrostatic charge for basic or acidic conditions, respectively (Figure 4). This charge shift results in weaker protein-protein hydrophobic

16

Jacek Jaczynski

interactions, while the protein-protein electrostatic repulsion becomes more predominant, resulting in protein-water interaction (i.e., water solubility). When proteins start interacting with water, a drastic increase of viscosity occurs. The viscosity drops as soon as the proteins become water soluble. However, this viscosity increase is an important processing parameter that may result in mixing issues (i.e., pH and protein solubility gradient), foam formation, and etc. unless the solution is continuously maintained at the desired pH as for example in a continuous protein and lipid recovery system (see below). While the muscle proteins are in full interaction with water (i.e., solubilized), the third step is applied. Typically, centrifugation separates the solution for light, medium, and heavy fractions containing fish oil, solubilized muscle proteins, and impurities (bones, skin, scale, skin, insoluble proteins, and etc.), respectively. While the hydrophobic triglycerides are relatively easy to separate from the solution, the membrane phospholipids are relatively persistent due to their amphiphilic characteristics (Figure 6). Although membrane phospholipids are present in smaller amounts in the fish muscle than triglycerides, the membrane phospholipids have been demonstrated to contribute more to rancidity development. Therefore, it is desirable to remove as much lipids as possible during the separation step. The third step results in separation of crude fish oil that is rich in ω-3 PUFA and can be further processed for numerous food and non-food application (Chen et al., 2007b). The heavy fraction is rich in minerals such as Ca, Mg, and P; and therefore, can be a main ingredient in the development of animal feeds as well as most likely more profitable pet foods (Chen et al., 2007b). The medium fraction, containing water-soluble fish muscle proteins is recovered and subjected to the second pH adjustment in step four. The pH is adjusted to the pI of the fish muscle proteins (pH = 5.5). At the pH = 5.5 fish muscle proteins precipitate due to increased protein-protein hydrophobic interactions and decreased protein-water interactions as well as decreased protein-protein electrostatic repulsion. Similar to the first pH adjustment in the second step, as the proteins gradually stop interacting with water dipoles, the viscosity increases significantly. This viscosity issue can be overcome by continuous maintenance of the pH at 5.5. The precipitated fish muscle proteins are separated from the process water typically by centrifugation. The muscle proteins retain their gel forming capability; and therefore, can be used as a functional and main ingredient in human foods applications such as for example surimi seafood (commonly referred to as imitation crab meat). The process water is typically as clear as the water added in the first step; and therefore, can be re-used if a continuous system is used. However, the purity of process water greatly depends on processing parameters. The temperature during all of the processing steps is typically controlled at 1-8°C in order to reduce protein and lipid degradation.

Fish Muscle Proteins and Lipids The solubility curves of fish muscle proteins as a function of pH are shown in Figure 8 (Chen and Jaczynski, 2007a). The major group of fish muscle proteins – myofibrillar proteins are water-soluble at acidic and basic pH. The minimum water solubility (i.e., maximum protein

Protein and Lipid Recovery from Food Processing By-Products…

17

precipitation) is between pH 5.0-6.0. However, ionic strength (IS) of the solution plays an important role in water solubility of fish myosin. When the IS was adjusted to 0.2, the minimum solubility of myofibrillar proteins shifted by approximately 1 pH unit towards precipitation at more acidic conditions (Chen and Jaczynski, 2007a). This difference is important if a continuous protein and lipid recovery system is used. If acid is used in the first pH adjustment (step 2), then base is required to adjust pH to 5.5 (step 4) (Figure 7). If proteins are solubilized at basic pH (step 2), then they are precipitated by acid addition (step 4). In either case, salt is generated in the reaction, which increases the IS. In the continuous recovery system the process water is re-cycled; and therefore, a salt accumulation occurs, affecting the pH at which myofibrillar proteins precipitate. Figure 8 also shows that second major group of fish muscle proteins – sarcoplasmic proteins are water-soluble at both acidic and basic pH. However, these proteins do not precipitate as demonstrably between pH 5.0-6.0 as the myofibrillar proteins. The IS also plays an important role in precipitation of the sarcoplasmic proteins. As the IS increases, the sarcoplasmic proteins precipitate more at pH = 5.5. Therefore, in a continuous system, the sarcoplamic proteins are recovered more as the salt accumulation increases. The isoelectric solubilization/precipitation results in relatively high recovery of proteins from the starting materials (Table 2). The data presented in Table 2 was determined for a lab bench-top system using a batch mode of operation in laboratory beakers and a typical laboratory stationary centrifuge. It is likely that the continuous system that re-cycles the process water would be even more efficient than the data presented in Table 2. In general, the protein recovery yields tend to be slightly higher when solubilization is carried out at acidic pH compared to the basic one (Table 2, Hultin et al., 2005). While the best recovery occurs when proteins are precipitated at pH = 5.5, the recovery drops more rapidly when proteins are precipitated at pH = 5.0 than pH = 6.0. Solubility of the myofibrillar proteins increases much more rapidly at acidic pH than at the basic one (Figure 8). Therefore, if precipiatation is conducted at pH = 5.0, the recovery yield is lower than at pH 6.0. Surimi processing is a commercial process that allows recovery of muscle proteins (primarily myofibrillar) from fish. Processing of 100 kg of life fish using surimi technology recovers about 19.5 kg of proteins (i.e., surimi) (Lee, 1999), while the isoelectric solubilization/precipitation allows about 90% protein recovery from the by-products (Table 2). The relatively low protein recovery by surimi technology means that the un-recovered proteins end up in the process water. In addition, surimi processing requires tremendous amounts of fresh water, about 20 times the weight of the deboned meat (Lee, 1999). The effluent process water is high in biological oxygen demand (BOD); and therefore, requires proper treatment before it can be discharged. Not only does the low recovery results in relatively low profitability of surimi technology, but also the treatment cost of high volumes of process water (or even plant closures due to environmental issues) are important factors associated with surimi technology.

(A)

0.20

Protein concentration in supernatant (g/L)

Protein concentration in supernatant (g/L)

2.5

Ionic strength (M NaCl) 0.16

2.0

0.12 1.5

0.08

1.0

Ionic strength (M NaCl)

18

Figure 8. Solubility of major fish muscle proteins, myofibrillar (A and B) and sarcoplasmic (C) proteins as a function of pH and ionic strength (IS). (A) IS was not adjusted, but monitored during isoelectric solubilization; (B) IS was adjusted to 0.20 with NaCl; (C) IS was only adjusted at pH = 5.50 (open circles). (Source: Adapted from Chen Jacek Jaczynski and Jaczynski (2007a)).

0.04

0.5

0.0

0.00 1

2

3

4

5

6

7

8

9

10

11

12

13

pH

(B)

Protein concentration in supernatant (g/L)

0.25

Ionic strength (M NaCl) 0.20

2.0

0.15

1.5

0.10

1.0

Ionic strength (M NaCl)

Protein concentration in supernatant (g/L)

2.5

0.05

0.5

0.0

0.00 1

2

3

4

5

6

7

8

9

10

11

12

13

pH

(C)

Solubility of water-soluble protein (%)

95

IS=0.01

85

IS=0.03 IS=0.05 IS=0.11

75

Ionic strength not adjusted

IS=0.14

Ionic strength adjusted

IS=0.17 IS=0.20 IS=0.51

65

IS=1.08

55

IS=2.91

45 1

2

3

4

5

6

7

8

9

10

11

12

13

pH

Figure 8. Solubility of major fish muscle proteins, myofibrillar (A and B) and sarcoplasmic (C) proteins as a function of pH and ionic strength (IS). (A) IS was not adjusted, but monitored during isoelectric solubilization; (B) IS was adjusted to 0.20 with NaCl; (C) IS was only adjusted at pH = 5.50 (open circles). (Source: Adapted from Chen and Jaczynski (2007a)).

Protein and Lipid Recovery from Food Processing By-Products…

19

Table 2. Protein recovery yields from trout processing by-products using isoelectric solubilization/precipitation (Source: Adapted from Chen and Jaczynski (2007b)) pH (solubilization/precipitation)

% protein recovery

2.5/5.5

89.0

2.5/5.0

81.9

2.5/6.0

85.9

2.0/5.5

91.3

3.0/5.5

86.2

12.5/5.5

84.4

12.5/5.0

77.7

12.5/6.0

83.4

12.0/5.5

82.9

13.0/5.5

88.1

Unlike surimi technology, the isoelectric solubilization/precipitation allows processing of by-products such as frames and heads or whole animals such as krill to recover functional proteins and lipids (Chen and Jaczynski, 2007a; 2007b). The first step in the isoelectric solubilization/precipitation is homogenization (i.e., grinding) of the by-products (Figure 7). Typically, a meat homogenizer such as for example Stephan Machinery MCH-10 (Stephan Machinery, Columbus, OH) is used to reduce particle size to below 0.2 mm in order to increase surface area for solubilization reaction in the subsequent step. Since the fillets are recovered during filleting of whole fish, the bone, skin, scale, etc. contents in the processing by-products are higher than those in whole fish. Therefore, if the by-products are used in the isoelectric solubilization/precipitation as a starting material and they are finely ground, it is important to determine where these impurities end up. The ash content is a good indicator of impurities associated with high content of bones, scales, skin, and etc. Therefore, determining ash content of a particular recovered fraction typically indicates where these impurities end up. Table 3 shows that the ash content (dry basis) of trout frames and whole krill is 13.91 and 17.36%, respectively. However, the ash content of boneless skinless trout fillets and krill tail meat is 5.54 and 11.09%, respectively. The isoelectric solubilization/precipitation results in the ash content at approximately 5 and 1.5% for krill and trout recovered proteins, respectively. Therefore, the recovered proteins using the isoelectric solubilization/precipitation likely contain fewer impurities (bone, skin, scale, etc.) than the boneless skinless fillets and tail meat (Chen et al., 2007b; 2007c).

20

Jacek Jaczynski

Table 3. Muscle proteins recovered from krill and trout processing by-products using Table 3. Muscle proteins recovered from krill and trout processing by-products using isoelectric solubilization/precipitation very ash content, suggesting that isoelectric solubilization/precipitation have have very low ashlow content, suggesting that impurities separatedininstep step 33 (Figure (Figure 7)7)retain ash-yielding components such as bone, impurities separated retain ash-yielding components such as bone, skin, scale, fins, etc. The % ash listed in both tables is on dry basis (Source: Adapted skin, scale, fins, etc. The % ash listed in (2007c)). both tables is on dry basis (Source: Adapted from Chen et al. (2007b); and Chen et al. from Chen et al. (2007b); and Chen et al. (2007c)) % Ash

Krill tail meat

11.09

Boneless skinless trout fillet

5.54

Whole krill

17.36

Trout frames (by-products)

13.91

2.0

5. 98

Recovered impurities

41.10

2.5

4. 32

3.0

4. 01

12.0

4. 88

12.5

5. 71

13.0

5. 74

Proteins solubilized at pH

Proteins solubilized at pH

% Ash

2.5

2.14

3.0

1.61

12.0

0.88

12.5

1.37

13.0

2.14

Table 3 also shows that the impurities recovered in step 3 as a heavy fraction (Figure 7) contain 41.10% of ash. This fraction is very high in important minerals such as Ca, P, Mg, etc.; and therefore, could be used in animal feeds and pet foods (Chen et al., 2007b; 2007c). Since the fish oil is removed from this fraction, unlike typical fishmeal, it should not impart a fishy (i.e., rancid) odor to the meat of animals fed this fraction. If the recovered proteins are to be used in human food products, knowing the nutritional value of these proteins will be essential. Adult humans are incapable of synthesizing nine amino acids (AA); and therefore, they have to be provided in the diet for proper development and functioning. This is why these nine AA are called essential AA (EAA). Table 4 compares EAA content of krill and trout proteins recovered with the isoelectric solubilization/precipitation (Chen et al., 2007b; 2007c). Table 4 also lists the EAA content for de-fatted soybean protein isolate and the EAA pattern for high-quality protein that meets human requirements as established by the Food and Nutrition Board (FNB) Research Council (Hui, 1999). The proteins recovered from trout processing by-products are highly nutritious and contain all of the EAA. Soybean proteins are a typical example of plant-derived proteins and while they are an excellent source of EAA, they seem lower in methionine when compared to the FNB requirements and also lysine is lower than in the animal-derived proteins. While trout frames and whole krill fall short in EAA when compared to the FNB pattern, these products are not used for direct human consumption without processing. The muscle proteins recovered from both trout by-products (i.e., frames) and whole krill are excellent sources of EAA. Methionine and lysine are abundant in both sources of the proteins. The muscle proteins recovered from krill are of excellent quality and could provide very good nutrition if used in human food products. Lysine concentration is also critical for certain non-food applications. The recovered proteins can be chemically modified to make bio-degradable super-absorbent hydrogel (SAH) (Damodaran, 2004). About 1 g of such a SAH is capable of trapping 400 g of water or saline

Protein and Lipid Recovery from Food Processing By-Products…

21

solution in a gel network. However, high concentration of lysine in the source of protein prior to the modification is essential. Currently, non-biodegradable hydrocarbon-based SAH are used in diapers, paper towels, etc.

Table 4. Muscle proteins recovered from krill and trout processing by-products using isoelectric solubilization/precipitation are rich in essential amino acids (EAA) (Source: Adapted from Chen et al. (2007b); and Chen et al. (2007c)). The content of EAA from soy protein isolate along with the EAA pattern for high-quality protein that meets human requirements as established by the Food and Nutrition Board (FNB) Research Council are listed for comparison purposes (Source: Adapted from Hui (1999)). Abbreviations: Thr – threonine, Val – valine, Met – methionine, Ile – isoleucine, Leu – leucine, Phe – phenylalanine, His – histidine, Lys – lysine, Trp – tryptophan

Krill proteins solubilized at pH

Trout proteins solubilized at pH

Essential Amino Acids Thr

Val

Met

Ile

Leu

Phe

Hi s

Lys

Trp

Total

Average

trout frames

1.8

2.2

1.4

1.8

3.1

1.6

1.2

3.5

0.5

17.2

17.2

2.0

3.7

4.6

2.6

3.9

6.6

3.4

2.1

7.4

1.0

35.3

2.5

3.4

4.3

2.2

3.6

6.0

3.1

1.9

6.7

0.9

32.3

3.0

3.7

4.7

2.6

4.0

6.6

3.4

2.1

7.3

0.9

35.2

12.0

3.8

5.0

2.6

4.2

6.9

3.5

2.3

7.6

1.1

37.2

12.5

3.9

4.9

2.6

4.1

6.9

3.5

2.2

7.6

1.1

36.9

13.0

4.1

5.1

2.6

4.3

7.1

3.7

2.3

7.8

1.2

38.2

whole krill

2.2

2.6

1.5

2.5

4.0

2.2

1.1

4.4

0.7

21.2

2.0

4.8

6.0

2.9

5.5

9.0

5.0

2.6

9.2

1.5

46.6

2.5

4.5

5.8

3.2

5.7

8.9

4.9

2.5

9.2

1.6

46.3

3.0

4.8

5.9

3.3

5.9

9.2

5.2

2.6

9.6

1.6

48.1

12.0

4.6

5.8

3.4

5.7

8.8

5.1

2.7

9.2

1.7

47.0

12.5

4.5

5.6

3.2

5.5

8.6

5.0

2.5

8.9

1.5

45.3

13.0

4.4

5.5

3.1

5.5

8.4

4.8

2.5

8.7

1.5

44.3

soybean

3.9

4.6

1.1

4.6

7.8

5.0

2.6

6.4

1.4

37.4

37.4

FNB

3.5

4.8

2.6

4.2

7.0

7.3

1.7

5.1

1.1

37.3

37.3

34.3

37.4

21.2

47.0

45.5

22

Jacek Jaczynski

Although solubilization at acidic pH generally results in higher recovery yields as compared to the basic solubilization (Table 2), the texture quality of protein gels made from proteins recovered at basic pH is better than that from proteins recovered at acidic pH (Figure 9, Hultin et al., 2005; Chen and Jaczynski, 2007a; 2007b).

Figure 9. In general, proteins solubilized at basic pH and subsequently recovered at pH 5.5 exhibit slightly better texture properties than the proteins solubilized at acidic pH. The proteins were recovered from trout processing by-products and whole krill (Source: Adapted from Chen and Jaczynski (2007a, 2007b)).

The gels made from proteins recovered at basic pH are firmer and have also been reported as whiter when compared to gels developed from proteins recovered at acidic pH (Hultin et al., 2005). The firmer texture suggest that the proteins recovered at basic pH retain higher quality than those recovered at acidic pH (Chen and Jaczynski, 2007a; 2007b). Therefore, despite slightly lower yields for basic solubilization (Table 2), the higher quality of recovered proteins may be more important for a processor. However, this particular factor will depend on a final application of the recovered proteins and price. Krill has extremely potent endogenous proteolytic enzymes (Kolakowski and Lachowicz, 1982), which in part technologically impeded a development of food products from krill (Suzuki, 1999; Tou et al., 2007). Figure 10 shows viscoelastic modulus (G’) of krill muscle proteins recovered using the isoelectric solubilization/precipitation. As the krill protein paste is subjected to slow heating ramp (1°C/min) in a dynamic rheometer, the proteins start gelling, which results in increased elasticity and decreased viscosity of the paste (increased G’). Beef plasma protein (BPP) has been used as a protease inhibitor in the surimi industry for surimi recovered from fish species prone to enzymatic proteolysis such as Pacific whiting. When krill protein paste was formulated without BPP and slowly heated in a dynamic rheometer, extensive proteolysis occurred up to 60°C and the proteins failed to form a gel. However, when 1% BPP (wt/wt) was added to the krill protein paste and subjected to the same heat ramp in the dymanic rheometer, the recovered proteins gelled (Chen and Jaczynski, 2007b) (Figure 10).

Protein and Lipid Recovery from Food Processing By-Products…

23

Figure 10. Muscle proteins recovered from krill using the isoelectric solubilization/precipitation showed poor gel-forming ability most likely due to high activity of endogenous proteases. However, application of beef plasma protein (BPP) inhibited proteolytic activity; and therefore, resulted in gelation of muscle proteins recovered from krill (Source: Adapted from Chen and Jaczynski (2007b)).

Therefore, it is likely that the krill proteases responsible for protein degradation are retained with the proteins during the isoelectric solubilization/precipitation similar to cathepsin L in Pacific whiting during water washing in surimi making (Choi et al., 2005). There are several protease inhibitors commercially available besides BPP. It is highly recommended to use an inhibitor in formulations with krill proteins whenever fast heating such as for example ohmic heating is unavailable and firm texture of a final product is desired. From an economic stand-point fast protein recovery is important. Protein separation during step 5 (Figure 7) is relatively slow due to small particle size of the muscle proteins that were first subjected to solubilization and subsequent precipitation. Therefore, relatively high g force and residence time in the decanter (step 5 in Figure 7) has to be allowed. The size of the protein particles can be increased by allowing protein-protein hydrophobic interactions to form over extended time (about 24 hr), following step 4 (pH adjustment to pH = 5.5). The temperature, however, has to be controlled at 1-8°C in order to minimize protein denaturation. Therefore, if time is extended, the solution of precipitated proteins should be stored under refrigeration. According to Stoke’s law (Equation 1), the particle settling velocity under gravitational force (g) is dependent on the following four variables: (1) density differential between phases, (2) viscosity, (3) “g” force, and (4) the square of particle size.

Equation (1) where: S – particle settling velocity ∆(ρ) – density differential between phases

24

Jacek Jaczynski g – gravitational force D – particle diameter µ – viscosity

The only variable in the isoelectric solubilization/precipitation recovery system that can be modified is variable (4). For example, if the particle size is increased by a factor of three, the particle settling velocity will increase by nine. Not only can the protein particle size be increased by hydrophobic protein-protein interaction over 24 hr, but also by addition of flocculants. Flocculants are commonly used by the food processing industry and in the treatment of drinking water for solution clarification purposes. There are many different flocculants commercially available. Figure 11 and 12 show protein separation following 10minute reaction with anionic flocculent of high molecular weight. The second pH adjustment (step 4 in Figure 7) typically requires 10 minutes in a continuous recovery system. This anionic flocculent can be injected into the bio-reactor (step 4) to induce protein flocculation; and therefore, increase separation efficiency in the subsequent step 5. This flocculent does not have an adverse effect on color or gelation properties of the recovered proteins. However, for commercial application, proper approvals would need to be obtained from local authorities. The optical density of the supernatant following 10-minute protein flocculation is comparable to that of clear water. Therefore, the process water can be re-cycled in step 1 (homogenization in Figure 7).

Figure 11. The particle size of fish muscle proteins precipitated in step 4 (Figure 7) can be efficiently increased by 10-minute reaction with anionic flocculent of high molecular weight. Therefore, subsequent separation in a decanter (step 5) can be performed at higher flow rates. The flocculent can be injected into bio-reactor (step 4). Optical density of the supernatant following the 10-minute reaction is comparable to that of clear water.

Protein and Lipid Recovery from Food Processing By-Products…

25

Figure 12. Protein flocculation can further enhance separation of the fish muscle proteins in the decanter (step 5 in Figure 7) following protein precipitation in step 4. On the left are precipitated fish muscle proteins without flocculent (control). On the right are muscle proteins following 10-minute reaction with commercially available flocculent. The optical density of the effluent process water (bottom of the beaker in the picture on the right) is comparable to that of clear water; and therefore, the process water can be re-used in the recovery system.

Figure 13. The ω-3 (n3) and ω-6 (n6) FA do not undergo significant degradation due to the recovery process. Therefore, these heart-friendly FA could be used in dietary supplements as well as functional foods (Source: Adapted from Chen et al. (2007b)).

Lipids are recovered in step 3 (Figure 7) as a light fraction. Fish lipids are rich in ω-3 PUFA such as DHA, EPA, and ALA. These FA are well recognized for their human health benefits are often used as ingredients in functional foods. However, at the same time these FA

26

Jacek Jaczynski

are particularly susceptible to oxidation. The isoelectric solubilization/precipitation applies relatively extreme pH, which could cause degradation of these FA. However, Figure 13 shows that the degradation of the most important fatty acids is minimal likely due to temperature control and short exposure time to the extreme pH during solubilization and first separation (steps 2 and 4 in Figure 7). Therefore, this crude fish oil could be used as a basis for further refining processes and subsequent application in numerous food (dietary supplements, functional foods, nutraceuticals, etc.) and non-food applications (cosmetics, industrial oils for paints, etc.).

Equipment Considerations A continuous meat homogenizer is used in step 1 (Figure 7). This homogenizer reduces particle size of the by-products to below 0.2 mm. This particle size results in a surface area that is great enough to result in efficient protein solubilization in step 2 as well as particle size large enough to allow efficient separation is step 3. Following homogenization, the homogenate is pumped to the bio-reactor one for 10-min solubilization reaction (Figure 14). The bio-reactor is equipped with a pH probe that continuously monitors pH of the solution in the vessel and feeds this pH to the control box.

Figure 14. Bio-reactors are capable of continuous and automatic pH adjustment, proper mixing, temperature control, continuous pumping in and out, as well as precise dosing of food-grade additives such as emulsion breakers, protein flocculants, and antifoam agents. Bio-reactor in the background is used for step 2 – protein solubilization (Figure 7), and bio-reactor in the foreground is used for step 4 – isoelectric precipitation. A control box is placed between the bio-reactors. Following step 2, lipids, protein solution, and insoluble impurities are separated in a decanter before the protein solution is pumped to the bio-reactor for precipitation (step 4). This set-up is working in a continuous mode at flow rate of 300 L/hr.

Protein and Lipid Recovery from Food Processing By-Products…

27

The control box is programmable and allows setting a pH value, which will be maintained in the vessel. As the incoming homogenate has a pH that is close to neutral (approximately 6.67.0), yet, the bio-reactor is programmed to maintain the vessel at pH = 11.00, a built-in pump will be triggered and a base will be rapidly pumped into the vessel to adjust pH to 11.00. Once the initial fill of the bio-reactor is finished, an equilibrium (homeostasis) is established and the bio-reactor will work continuously at the programmed pH. The bio-reactors in the continuous protein and lipid recovery system are also equipped with mixing baffles that allow gentle mixing to prevent pH gradient and excessive air intake that could cause foaming. Temperature in the vessel is controlled in a similar fashion as the pH, where the temperature probe can trigger flow of refrigerant to maintain programmed temperature. Not only are the small built-in pumps used for acid/base flow, but they also allow control of foam formation by injection of proper food-grade antifoam agents (by mechanism similar to the pH and temperature control) as well as emulsion breakers. The emulsion breaking can result in greater removal of fish oil, particularly persistent fish membrane lipids, from the solution following step 2 (Figure 7). The bio-reactors are connected to pumps that control flow rate through the system. The recovery system shown in Figure 14 can work at 300 L/hr, resulting in processing capability of about 43 kg of starting material per hour. Although these small scale bio-reactors are commercially manufactured from glass and stainless steel components, industrial strength polyethylene (PE) can easily withstand pH values used in the isoelectric solubilization/precipitation and harsh conditions in the fish processing industry. Based on the tested design of the bio-reactor system shown in Figure 14, a modular 150-gal bio-reactor system with a flow rate of 15 GPM (gallons per minute) has also been designed for processing of 11 tons of by-products per day (Figure 15). The use of PE in this modular system is very competitive and likely justifiable at industrial scale. If more processing capacity is required, either modules can be connected or larger vessels manufactured. While the bio-reactor is step 2 (Figure 7) works in rather extreme pH (either acidic or basic), the bio-reactor is step 4 works under relatively mild conditions (pH = 5.50). Therefore, special precautions should be taken when designing a bio-reactor to be used for protein solubilization in step 2. It needs to be pointed out that typically concentrated acids and bases are used with this technique. Therefore, special attention should be paid for these hazardous chemicals. Following 10-minute pH adjustment in step 2, the solution is pumped to a decanter for separation. A general view and a cross-sectional view are shown in Figure 16. Industrial decanters typically achieve a g force below 4,000 x g. While the decanter in step 3 works under extreme pH conditions, the decanter in step 5 works under rather mild pH conditions (pH = 5.50). Therefore, if leaks occur in the step 3, special safety steps and containment procedures should be devised prior to running this equipment. If the recovery system works in a continuous mode, the flow rates of all the steps should match. Therefore, if the discharge from the bioreactors is 300 L/hr, the decanters should be capable of handling the same flow. Otherwise, overflowing results and the personnel may be exposed to acid/base hazard. Decanters are commonly available in surimi processing plants. However, surimi technology does not work under acidic or basic pH conditions; therefore, testing of the decanters for application in the

28

Jacek Jaczynski

isoelectric solubilization/precipitation should be performed prior to the use at industrial scale. Decanters of various sizes are commercially available from few manufacturers.

Figure 15. A modular bio-reactor system for processing of 11 tons of fish by-products per day. This system is skid-mounted for portability.

Protein and Lipid Recovery from Food Processing By-Products…

29

Figure 16. Decanter centrifuges are typical separation equipment in the fish processing industry and can be employed for separation purposes in the continuous protein and lipid recovery system using the isoelectric solubilization/precipitation at a pilot/commercial scale. A commercial unit in shown on the top and a cross-sectional view of a decanter centrifuge bowl is shown on the bottom (courtesy of Alfa Laval).

Figure 17 shows the materials that can be recovered from fish processing by-products. In step 3 (Figure 7), fish oil rich in ω-3 PUFA is recovered along with the impurities that are rich in several important minerals (Ca, Mg, P, etc.). In step 5, the functional muscle proteins are recovered along with water that is re-cycled in step 1.

Figure 17. Three major materials are recovered from fish processing by-products using isoelectric solubilization/precipitation: (1) fish oil recovered in step 3 (Figure 7), (2) impurities (bone, skin, scale, fin, insoluble proteins, etc.) recovered in step 3, (3) functional muscle proteins that retain their gelforming ability; and therefore, (4) value-added human foods can be developed from these recovered protein. The muscle proteins are recovered in step 5. Following protein separation in step 5, water (not shown) is used again (re-cycled) in step 1.

30

Jacek Jaczynski

Conclusion The amount of by-products generated from processing of aquatic foods as well as the bycatch, discards, and low-value fishes is staggering. Yet, at the same time, human kind has already started over-fishing of several species. Aqauculture has experienced great growth during the past 25 years, filling the supply gap that cannot be filled by our dwindling natural resource. However, fish processing has not changed significantly, often contributing to increased pollution. This presents a great opportunity as well as a challenge. Isoelectric solubilization/precipitation of fish muscle proteins has long been applied to recover milk and soy proteins. Hopefully, by thorough understanding of fish muscle proteins and their isoelectric behavior we will be able to develop this technology at a commercial scale for efficient fish processing. However, for a commercial success, it will be necessary to develop final applications for the recovered materials. Probably similar model to the diary and soybean processing industries could be used for the development of final food products. Government agencies as well as academia should be actively engaged in the collaborative development of final food products.

References Anonymous. 2004. The state of world fisheries and aquaculture. Food and Agriculture Organization. Rome (Italy). Chen, Y.C., Jaczynski, J. 2007a. Protein recovery from rainbow trout (Oncorhynchus mykiss) processing by-products via isoelectric solubilization / precipitation and its gelation properties as affected by functional additives. Journal of Agricultural and Food Chemistry 55:9079-9088. Chen, Y.C., Jaczynski, J. 2007b. Gelation of protein recovered from whole Antarctic krill (Euphausia superba) by isoelectric solubilization / precipitation as affected by functional additives. Journal of Agricultural and Food Chemistry 55:1814-1822. Chen, Y.C., Nguyen, J., Semmens, K., Beamer, S., Jaczynski, J. 2007a. Physicochemical changes in omega-3-enhanced farmed rainbow trout (Oncorhynchus mykiss) muscle during refrigerated storage. Food Chemistry 104(3):1143-1152. Chen, Y.C., Nguyen, J., Semmens, K., Beamer, S., Jaczynski, J. 2006. Enhancement of omega3 fatty acid content in rainbow trout (Oncorhynchus mykiss) fillets. Journal of Food Science 71(7):C383-389. Chen, Y.C., Tou, J.C., Jaczynski, J. 2007b. Amino acid, fatty acid, and mineral profiles of materials recovered from rainbow trout (Oncorhynchus mykiss) processing by-products using isoelectric solubilization / precipitation. Journal of Food Science 72(9):C527-535. Chen, Y.C., Tou, J.C., Jaczynski, J. 2007c. Composition and recovery yield of protein and other components isolated from whole Antarctic krill (Euphausia superba) by isoelectric solubilization / precipitation. Journal of Food Composition and Analysis. Under review. Choi, Y.J., Kang, I.S., Lanier, T.C. 2005. Proteolytic enzymes and control in surimi. In: Surimi and Surimi Seafood, 2nd ed. Park, J.W. (ed.). CRC Press. Boca Raton (United States). Damodaran, S. 2004. Protein-polysaccharide hybrid hydrogels. U.S. Patent No. 6,821,331.

Protein and Lipid Recovery from Food Processing By-Products…

31

Gildgerg, A. 2002. Enhancing returns from greater utilization. In: Safety and Quality Issues in Fish Processing. Bremner, H.A. (ed.). Woodhead Publishing Limited. Cambridge (U.K.). Hui, Y.H. 1999. Soybean and soybean processing. In: Wiley Encyclopedia of Food Science and Technology (2nd ed.). Francis, F.J. (ed.). John Wiley and Sons. Hoboken (United States). Hultin, H.O., Kristinsson, H.G., Lanier, T.C., Park, J.W. 2005. Process for recovery of functional proteins by pH shift. In: Surimi and Surimi Seafood, 2nd ed. Park, J.W. (ed.). CRC Press. Boca Raton (United States). Hultin, H.O., Kelleher, S.D. 2002. Protein composition and process for isolating a protein composition from a muscle source. U.S. Patent No. 6,451,975. Hultin, H.O., Kelleher, S.D. 2001. Process for isolating a protein composition from a muscle source and protein composition. U.S. Patent No. 6,288,216. Hultin, H.O., Kelleher, S.D. 2000. High efficiency alkaline protein extraction. U.S. Patent No. 6,136,959. Hultin, H.O., Kelleher, S.D. 1999. Process for isolating a protein composition from a muscle source and protein composition. U.S. Patent No. 6,005,073. Kolakowski, E., Lachowicz, K. 1982. Application of partial autoproteolysis to extraction of protein from Antarctic krill (Euphausia superba). Part 3. Changes in and yield of nitrogen substances during autoproteolysis of fresh and frozen krill. Die Nahrung 26: 933-939. Kreuzer, R. 1974. Fish and its place in culture. In: Fishery Products. Kretizer, R. (ed.). Fishing News (Books). West Byfleet (U.K.). Lanier, T.C., Carvajal, P., Yongsawatdigul, J. 2005. Surimi gelation chemistry. In: Surimi and Surimi Seafood, 2nd ed. Park, J.W. (ed.). CRC Press. Boca Raton (United States). Lee, C.M. 1999. Surimi: Science and technology. In: Wiley Encyclopedia of Food Science and Technology (2nd ed.). Francis, F.J. (ed.). John Wiley and Sons. Hoboken (United States). McCann, A.M. 1988. The Roman port of Cosa. Scientific American 258(3):84-91. Meinke, W.W., Rahman, M.A., Mattil, K.F. 1972. Some factors influencing the production of protein isolates from whole fish. Journal of Food Science 37:195-198. Meinke, W.W., Mattil, K.F. 1973. Autolysis as a factor in the production of protein isolates from whole fish. Journal of Food Science 38:864-866. Murano, P.S. 2003. Understanding Food Science and Technology. Wadsworth/Thomason Learning. Belmont (United States). Nicol, S., Endo, Y. 1999. Krill fisheries: Development, management and ecosystem implications. Aquatic Living Resources 12(2):105-120. Raa, J., Gildberg, A. 1982. Fish Silage: A Review. CRC Press. Boca Raton (United States). Stewart, K.M. 1994. Early hominid utilization of fish resources and implications for seasonality and behaviour. Journal of Human Evolution 27(13):229-245. Strom, T., Eggum, B.O. 1981. Nutritional value of fish viscera silage. Journal of the Science of Food and Agriculture 32:115-120. Suzuki, T. 1999. Krill protein processing. In: Wiley Encyclopedia of Food Science and Technology (2nd ed.). Francis, F.J. (ed.). John Wiley and Sons. Hoboken (United States). Suzuki, T. 1981. Fish and Krill Protein: Processing Technology. Applied Science Publishers. London (U.K.).

32

Jacek Jaczynski

Suzuki, T., Shibata, N. 1990. The utilization of Antarctic krill for human food. Food Reviews International 6(1): 119-147. Tou, J.C., Jaczynski, J., Chen, Y.C. 2007. Krill for human consumption: nutritional value and potential health benefits. Nutrition Reviews 65(2):63-77. Vannuccini, S. 2004. Overview of fish production, utilization, consumption and trade. Food and Agriculture Organization Fisheries Information, Data and Statistics Unit. Rome (Italy). Zugarramurdi, A., Parin, M.A., Lupin, H.M. 1995. Economic engineering applied to the fishery industry. Food and Agriculture Organization Fisheries Technical Paper 351. Rome (Italy). Whitney, E., Rolfes, S.R. 2005. Understanding Nutrition, 10th ed. Wadsworth/Thomason Learning. Belmont (United States).