Structured Lipids Novel Fats with Medical, Nutraceutical, and Food Applications

Transcription

1 Structured Lipids Novel Fats with Medical, Nutraceutical, and Food Applications H.T. Osborn and C.C. Akoh ABSTRACT CT: : Gener enerally ally,, structur uctured lipids (SLs) are triacylgly iacylglycer cerols (TAGs) that have been modified to change the fatty acid composition and/or their positional distribution in glycerol backbone by chemically and/or enzymatically cata- lyzed reactions and/or genetic engineering. ing. Mor ore specifically,, SLs are modified TAGs with improved nutritional or functional properties. SLs provide an effective means for producing tailor-made lipids with desired physical character- istics, chemical properties, and/or nutritional benefits. The production, commercialization outlook, medical, and food applications of SLs are reviewed here. Physical property measurements for SL in food systems and future research needs for increased industrial ial acceptance are also included in this review eview. Keywords: acidolysis, interesterification, lipases, modified fats, structured lipids Introduction With the ability to combine the beneficial characteristics of component fatty acids into 1 triacylglycerol molecule, lipid modification enhances the role fats and oils play in food, nutrition, and health applications. Structured lipids are tailor-made fats and oils with improved nutritional or physical properties because of modifications to incorporate new fatty acids or to change the position of existing fatty acids on the glycerol backbone. Soybean and safflower oils have been used in making fat emulsions for total parenteral nutrition (TPN) and enteral administration. Later, a physical blend of medium chain triacylglycerols (MCTs) and long chain triacylglycerols (LCTs) was used, with the MCTs being readily metabolized for quick energy. More recently, SLs were designed to provide simultaneous delivery of beneficial long chain fatty acids (LCFAs) at a slower rate and medium chain fatty acids (MCFAs) at a quicker rate (Babayan 1987; Akoh 1998). SL synthesis yields novel triacylglycerol (TAG) molecules (Akoh 1998). SLs may provide the most effective means of delivering desired fatty acids for nutritive or therapeutic purposes, and for targeting specific diseases and metabolic conditions (Lee and Akoh 1998). Improvements or changes in the physical and/or chemical characteristics of a TAG can also be achieved when SLs are synthesized. Lipid modification strategies for the production of functional fats and oils include chemically- or lipase-catalyzed interesterification and/or acidolysis reactions and genetic engineering of oilseed crops. Interesterification is used to produce fats with desirable functional and physical properties for food applications. Interest in interesterification from nutritional and functional standpoints is on the rise because of the possibility to produce transfree margarines, cocoa butter substitutes, and reduced calorie foods; to improve functional and physical properties of foods; and to improve the nutritional quality of fats and oils. Regardless of the method used, all SLs are designed to incorporate various fatty acids based on their potential benefits for improvements in nutrition or physical characteristics. This review covers a broad range of information concerning the production of SLs, including component fatty acids, production strategies, medical and nutraceutical applications, functional food applications, commercialization outlook, and future prospects for research and development in this field. Component Fatty Acids A variety of fatty acids are used in the synthesis of SLs, taking advantage of the functions and properties of each to maximize the benefits of a given SL. The component fatty acids and their position in the TAG molecule determine the functional and physical properties, the metabolic fate, and the health benefits of an SL. Therefore, careful analysis of the function and metabolism of component fatty acids is merited. Short chain fatty acids Short chain fatty acids (SCFAs) range from 2 to 6 carbons long, and are also known as volatile fatty acids. Traditional sources of SCFAs include bovine milk and butter fat. Due to their water soluble nature, molecular size, and short chain length, they are more rapidly absorbed in the stomach than other fatty acids. Additionally, SCFAs attached to the sn-3 position of TAGs will be completely hydrolyzed in the lumen of the stomach and small intestine, due to the positional and chain length specificity of human pancreatic gastric lipase. SCFAs have lower heat of combustion than other fatty acids, making them lower in calories. Caloric values of common SCFAs are as follows: C2:0, 3.5 kcal/g; C3:0, 5.0 kcal/g; C4:0, 6.0 kcal/g; and C6:0, 7.5 kcal/g (Akoh 1998). Medium chain fatty acids The primary sources of MCFAs, which range in length from 110 COMPREHENSIVE REVIEWS IN FOOD SCIENCE AND FOOD SAFETY Vol. 3, Institute of Food Technologists

2 C6:0 to C12:0, are coconut and palm kernel oils. MCFAs are preferentially transported via the portal vein to the liver, because of their smaller size and greater solubility compared to LCFAs (Bell and others 1991; Straarup and Hoy 2000). For entry into the mitochondria of all tissues, MCFAs are not carnitine-dependent (Babayan 1987; Bell and others 1997). Additionally, MCFAs are metabolized as rapidly as glucose in the body with little tendency to deposit as stored fat, because they are not readily re-esterified into triacylglycerols. Although MCFAs may be useful in the control of obesity, they can potentially raise serum cholesterol levels. Therefore, it appears that MCFAs are most useful in a structured lipid that combines their inherent mobility, solubility, and ease of metabolism with more healthful polyunsaturated fatty acids (Cater and others 1997; Akoh 1998). Long chain fatty acids Long chain fatty acids (LCFAs), ranging from C14 to C24, are common to animal fats and vegetable and marine oils. LCFAs are absorbed and metabolized more slowly than either medium or short chain fatty acids; much of the LCFAs may be lost as calcium-fatty acid soap in the feces (Broun and others 1999). LCFAs cannot be absorbed or transported in the blood, due to their increased hydrophobic character compared to SCFA and MCFA. Instead they must be first packaged into micelles, and then enter the intestinal cells where chylomicrons are formed. Chylomicrons are secreted into the lymphatic system and ultimately enter the systemic circulation. Carnitine is then required to transport LCFAs into the mitochondria of cells. Several types of LCFAs exist, and the important ones in SL production will be discussed here. Omega-6 fatty acids cannot be synthesized by humans and are therefore considered essential fatty acids (EFAs). Linoleic acid (18:2n-6), found in most vegetable oils and plant seeds, is an EFA that can be desaturated further, and elongated to arachidonic acid (20:4n-6). Arachidonic acid is a precursor for eicosanoid formation. Another type of EFA is the omega-3 fatty acid, such as linolenic acid (18:3n-3), which is found in soybean and linseed oils. Eicosapentaenoic acid, 20:5n- 3 (EPA), and docosahexaenoic acid, 22:6n-3 (DHA), found in fish oil, are other n-3 polyunsaturated fatty acids (PUFAs) of interest in SL production. The n-3 fatty acids are essential in growth and development throughout the human life cycle and should be included in the diet. The n-3 PUFAs inhibit tissue eicosanoid biosynthesis and reduce inflammation. Diets rich in n-3 PUFAs also increase high density lipoprotein (HDL) cholesterol, while decreasing low density lipoprotein (LDL) and very low density lipoprotein (VLDL) cholesterol levels. Omega-9 fatty acids, found as oleic acid (18:1n-9) in many vegetable oils, are not EFAs, but play a moderate role in reducing plasma cholesterol in the body. Conjugated linoleic acid (CLA), which has been shown to exhibit potent anticancer properties in animal models of carcinogenesis, can also be used for designing SLs. The major dietary source of this important class of PUFA is from the meats and fats of ruminant animals. Commercial sources of foodgrade quantities of this unusual non-methylene-interrupted fatty acid will be required before the therapeutic benefits are realized in humans (Watkins and German 1998). Long chain saturated fatty acids are generally believed to increase serum cholesterol levels. However, stearic acid (18:0) is neutral with respect to cholesterol levels in the blood, partly because it has a melting point that is higher than body temperature and it is readily desaturated to oleic acid in vivo. Additionally, TAGs with high amounts of long chain saturated fatty acids are poorly absorbed in humans. Controlling the positional distribution of component fatty acids in the final TAG is also important. During digestion, TAGs are degraded to sn-2 monoacylglycerols (MAGs) and free fatty acids in the small intestine by pancreatic lipase. The sn-2 MAG and free fatty acids are then absorbed by the enterocytes. In the intestinal mucosa cells, sn-2 MAGs are re-esterified with fatty acids of exogenous or endogenous origin to form new TAGs. These are then packed into chylomicrons and excreted into the lymph. The rate of hydrolysis of TAGs by pancreatic lipase is affected by chain length and unsaturation of the fatty acids at the sn-1 and -3 positions, with medium chain triacylglycerols (MCTs) being degraded faster than long chain triacylglycerols (LCTs) (Straarup and Hoy 2000). Production of Structured Lipids Chemical synthesis Chemical interesterification is inexpensive and easy to scale up; however, the reaction lacks specificity and offers little or no control over the positional distribution of fatty acids in the final product (Willis and Marangoni 1999). This process usually involves hydrolysis of a mixture of MCTs and LCTs, and re-esterification after random mixing of the MCFAs and LCFAs has occurred, by the ester-interchange reaction (Akoh 1998; Lee and Akoh 1998). Ester-interchange: R 1 -CO-OR 2 + R 3 -CO-OR 4 R 1 -CO-OR 4 +R 3 -CO-OR 2 This reaction, catalyzed by alkali metals or metal alkylates, requires high temperatures and anhydrous conditions. In addition to the desired randomized TAGs, a number of unwanted products are also obtained from this reaction and may be difficult to remove (Akoh 1998). Enzymatic synthesis Lipases occur widely in nature and are active at the oil-water interface in heterogeneous reaction systems. Lipase catalyzed interesterification reactions offer the advantage of greater control over the positional distribution of fatty acids in the final product, due to lipases fatty acid selectivity and regiospecificity. Lipases hydrolyze TAGs to monoacylglycerols, diacylglycerols (DAGs), free fatty acids (FFA), and glycerol. In addition to the ester-interchange reaction discussed in the previous section, lipases can also catalyze direct esterification, acidolysis, and alcoholysis reactions (Lee and Akoh 1998). Direct esterification: R 1 -CO-OH + R-OH R 1 -CO-OR + H 2 0 Acidolysis: R 1 -CO-OR + R 2 -CO-OH R 2 -CO-OR + R 1 -CO-OH Alcoholysis: R-CO-OR 1 + R 2 -OH R-CO-OR 2 + R 1 -OH Lipase-catalyzed reactions are a combination of esterification and hydrolysis (reverse reaction) reactions. Water must be continuously removed from the reaction medium in order to increase esterification reactions, while minimizing hydrolysis in order to obtain high conversion rates to products. When excess water is present, hydrolysis predominates, resulting in the accumulation of glycerol, FFAs, MAGs, and DAGs. However, some water is essential for enzymatic catalysis because it maintains enzyme dynamics during noncovalent interactions. To optimize 1-step lipasecatalyzed reactions, it is necessary to strike a balance between hydrolysis and esterification. Willis and Marangoni (1999) recently proposed splitting the reactions into 2 component hydrolytic and esterification phases and then optimizing the reaction conditions for each individual phase, in order to improve productivity and potentially make lipase-catalyzed reactions economically viable. The results of their study indicated that using chemical esterification of oil for the 1st phase of the reaction and enzyme hydrolysis in the 2nd phase offered the best control over the fatty Vol. 3, 2002 COMPREHENSIVE REVIEWS IN FOOD SCIENCE AND FOOD SAFETY 111

3 CRFSFS: Structured lipids. Comprehensive.. Reviews in Food Science and Food Safety acid positional distribution in the final product and required the least amount of reaction time. There were drawbacks associated with the chemical reaction, and further studies are needed in order to reduce product charring, degree of hydrolysis, and decomposition. Structured lipids can be produced with lipases in organic solvent, where substrates are soluble and hydrolysis can be minimized. The type of organic solvent employed can dramatically affect the reaction kinetics and catalytic efficiency of an enzyme. The extent to which the solvent affects the activity or stability of the enzyme and the effect of the solvent on the equilibrium position of the desired reaction must both be considered when choosing a solvent for biocatalysis. Hydrophilic or polar solvents can penetrate into the hydrophilic core of proteins and alter their functional structure. They also strip off the essential water of the enzyme. Hydrophobic solvents are less likely to cause enzyme inactivation in esterification reactions (Akoh 1998). Most lipases are optimally active between 30 and 40 C (Shahani 1975). As the temperature increases, enzyme molecules unfold by destruction of bonds, such as sulfide bridges, and may lead to hydrolysis of peptide bonds and deamidation of asparagine and glutamine residues. However, these processes can be avoided in a water-free environment. Immobilization of enzymes also results in greater thermostability. Additionally, genetically engineered lipases are now available for the synthesis of SLs. It is hoped that the use of biotechnology will reduce the cost of lipases, making the enzymatic route to SLs economically viable. Other factors affecting enzymatic activity and product yield include ph, substrate molar ratio, enzyme activity and load, incubation time, specificity of enzyme to substrate type and chain length, and regiospecificity (Akoh 1998). Two of the most attractive reasons for choosing enzymatic over chemically catalyzed reactions for SL production are the energy saved and minimization of thermal degradation. Genetic engineering of plant lipids Limited fatty acid compositional modifications of crop plant oils have been achieved in the past through traditional breeding techniques. Breeders have made use of the natural diversity that exists among plant varieties to transfer desirable characteristics from one to another. Mutagenesis has been used to produce cultivars of sunflower containing 90% oleate and less than 7% saturates. Sunflower oil normally contains only 16 to 20% oleate. This high-oleate sunflower is now in production and does not require catalytic hydrogenation for stabilization. Mutagenesis also enabled the production of flax that contains no linolenic acid, but has an increased amount of linoleic acid. Recently, high stearate soybean and high oleic sunflower oils with improved stability have been produced. Low saturate and linolenic acid soybean oils are desirable outcomes of genetic engineering. Most of the genes relevant to the synthesis of plant storage lipids have now been isolated. Genetic codes are available to introduce double bonds, elongate carbon chains, synthesize eicosapentaenate, and produce fatty acid isomers not normally found in common sources of edible oils. Plant engineers are now trying to incorporate the principles used in chemical and enzymatic synthesis of tailor-made structured lipids into their genetic engineering techniques. Researchers have targeted specific traits for incorporation into oilseed crops. The temperate zone oilseed crops tend to have storage lipids rich in unsaturated fatty acids, and thus liquid oils are, in themselves, not suitable for products like shortenings and margarines. Both palmitic and stearic acids are precursors to the unsaturated fatty acids, making up the bulk of the oil found in seeds from temperate zone crops. Pathways have now been engineered that produce fewer unsaturated fatty acids and accumulate more of the precursor palmitic and stearic acids in the TAGs. PUFAs have beneficial effects on serum cholesterol, but are highly susceptible to oxidation when not protected by antioxidants. Since oleic acid (18:1) appears to have a similar effect on cholesterol as linoleic acid (18:2n-6) and is not as susceptible to oxidation, researchers increased the ratio of monounsaturated fatty acids (MUFAs) to PUFAs in soybean and canola oil by modifying the activity of a microsomal membrane-bound oleate desaturase (Broun and others 1999). The presence of trans fatty acids in the diet has recently become a major health concern for consumers. Some well-publicized studies have suggested that trans fatty acids contribute to coronary heart disease by raising levels of LDL cholesterol and lowering levels of HDL cholesterol (Kris-Etherton 1995). Trans fatty acids are produced during the hydrogenation process used by food companies that try to produce a more solid fat from vegetable oils for use in shortenings and margarines. Several companies are actively pursuing the development of seed oils that contain levels of saturated fatty acids high enough to permit the elimination of the need for hydrogenation, and, subsequently, the production of trans fatty acids (Knauf and Del Vecchio 1998). Cloning and characterizing genes for a family of thioesterases was the 1st step toward the goal of incorporating MCFAs into oil seed crops that naturally do not contain such fatty acids. A gene from the California bay tree that produces MCFAs in its seeds was incorporated into canola plants. The transgenic canola now accumulates up to 65% more lauric acid in their seed TAGs (Voelker and others 1996). The sn-2 acyltransferase has a high degree of specificity for an unsaturated fatty acid; therefore, most of the oleic acid found in these TAGs is at the sn-2 position. This oil was commercialized by Calgene (now Monsanto) and marketed as Laurical. Initial functional screening of this new oil was conducted to see if it would serve as a cocoa butter replacement in coating and confectionery products. When evaluated in a standard confectionery coating versus commercial lauric fats (palm kernel oil and coconut oil), significant improvements were observed, including better compatibility with cocoa butter, significant increases in flavor impact, increased shelf life because of decreased bloom, and preferred mouthfeel. Also, some negative differences were found, including problems with demolding, less snap, and less gloss (Knauf and Del Vecchio 1998). This oil can also be used in coffee whiteners, whipped toppings, and filling fats (Broun and others 1999). Unfortunately, Laurical is not used much in foods, probably because lauric acid has the tendency to raise serum cholesterol levels. Laurical may find more use in nonfood products. Other thioesterase genes have been isolated that encode C8 and C10 fatty acids. Rapeseed plants expressing this gene accumulated significant amounts of these medium chain fatty acids (Dehesh and others 1996). It seems likely that producing a wide variety of MCFA will soon be possible in canola and other wild and domesticated Cuphea plant varieties. One problem observed in the above-mentioned transgenic plants is that MCFAs are excluded from the sn-2 position of the TAG. However, the genes encoding for lysophosphatidic acid sn-2 acyltransferases were recently isolated and may prove useful for producing nutritionally useful structured TAG molecules (Broun and others 1999). With the identification and purification of so many genes involved in plant seed lipid production, the possibilities for genetically engineered SLs may seem limitless. However, there are some limits to what is practically possible, and feasibility is the 1st consideration when thinking of specific ways to modify seed storage lipids (Knauf and Del Vecchio 1998). Lipids cannot be modified in such a way that they interfere with the ability of seeds to germinate or metabolize energy. A modified seed lipid will be useless if the crop is not viable. When unusual fatty acids are in- 112 COMPREHENSIVE REVIEWS IN FOOD SCIENCE AND FOOD SAFETY Vol. 3, 2002

4 corporated into the storage lipids during seed development, it is highly likely that they will also become incorporated into the structural lipids because their synthesis pathways share many common steps. The structural lipids are essential to support growth and for normal seed function. Therefore, re-engineering TAG biosynthesis must not compromise the synthesis of the structural lipids necessary to support growth and viability. Another type of limitation is the amount of oil obtainable from a crop, since most of the oilseed crop breeders measure success by how much seed is harvested per acre. Therefore, designer plants cannot compromise crop yield. Practical limitations also exist in this new field of genetic engineering. Expertise from a variety of fields is required to successfully produce genetically modified crops: plant physiology, enzymology, molecular biology, gene transfer cell biology, prototype evaluation, and plant breeding. Financial modeling and analyses of the eventual value and return, versus development costs and risk, may argue against ever starting certain projects. A major concern associated with genetically modified crops is the potential for cross-pollination to occur among genetically modified crops and unmodified plants in neighboring fields. This phenomenon would reduce the purity of both crops. Regulatory and political issues are also major obstacles facing genetic engineering. Cost of raw materials and synthetic capacities are factors to consider when this approach is compared with chemically or enzymatically modified lipids. Genetically engineered crops are limited by the number of acres that can be planted, and, in the latter case, huge capital investments are required to form manufacturing plants (Knauf and Del Vecchio 1998). Medical and Nutraceutical Applications of Structured Lipids The relationship between stereospecific fatty acid location and lipid nutrition suggests that the process of interesterification, or acidolysis, could be used to improve the nutrition profile of certain TAGs. Manufacturers of specialty food ingredients for infant formulas and enteral supplements should design fats with saturated fatty acids at the sn-2 position to provide increased caloric intake (Decker 1996). Infant formulas Ideally, the fat component of infant formulas should contain the fatty acids, such as MCFAs, linoleic acid, linolenic acid, and PU- FAs in the same position and amount as those found in human milk. Human milk is comprised of 20 to 30% palmitic acid, with 33% at the sn-2 position (Willis and others 1998). The fat in most infant formulas is of vegetable origin and tends to have unsaturated fatty acids in the sn-2 position. Innis and others (1994) found that infants fed human milk had 26% palmitic acid in their plasma TAGs, compared with 7.4% in infants fed vegetable oil-based infant formula with the same total concentration of palmitic acid, but not at the sn-2 position. When rats were fed a coconut oil and palm olein SL, absorption was increased due to the increased proportion of long-chain saturates at the sn-2 position (Lien and others 1993). Therefore, SLs with high proportions of palmitic acid at the sn-2 position would provide a fat with improved absorption capability in infants (Willis and others 1998). Care must be taken with regard to the concentrations of the saturates at the sn-2 position, because palmitic acid is the only saturate that has been studied extensively, and other long chain saturates may have hypercholesterolemic effects (Pai and Yeh 1997). Enteral and parenteral nutrition While MCTs have many advantages, a minimum amount of LCT is still necessary to provide EFAs. Physical mixtures of MCTs and LCTs have proven useful in the past for enteral (oral tube feeding) and parenteral (intravenous feeding) nutrition. More recently, structured TAGs comprised of LCFAs and MCFAs have emerged as the preferred alternative to physical mixtures for treatment of patients, although both products provide identical fat contents. Structured lipids comprised of both LCFAs and MCFAs are designed to provide simultaneous delivery of the fatty acids and a slower, more controlled release of the MCFA into the bloodstream (Babayan 1987). The advantages of enterally fed SLs may well relate to differences in absorption and processing. Structured TAGs that contain MCFA may provide a vehicle for rapid hydrolysis and absorption, due to their smaller molecular size and greater water solubility in comparison to long-chain TAGs (Jensen and Jensen 1992). Structured lipids offer several advantages over native oils and physical mixtures, including improved immune function, decreased cancer risk, thrombosis prevention, cholesterol lowering, improved nitrogen balance, and no risk of reticuloendothelial system (RES) impairment (Kennedy 1991). Increases in protein synthetic rates in both skeletal muscle and liver have also been demonstrated in patients receiving SL (DeMichele and others 1988). The more recent studies on SL for medical applications are an attempt to define the factors affecting the absorption and metabolic fates of a structured TAG, and the effects they have in vivo. The results of several such studies are summarized below, and the TAG structure of the SL in each study is diagrammed in Table 1. Lee and others (2000) showed that although they have a similar total fatty acid composition, enzymatically modified lipids and a physical mixture of lipids have different metabolic pathways, based on the structure and due to the difference in fatty acid composition at the sn-2 position of lipid molecules. In this study, soybean oil was modified by incorporating caprylic acid (C8:0) at the sn-1 and sn-3 positions; the LCFA remained intact at sn-2 position. This SL was compared to a physical mixture of soybean oil and tricaprylin. Caprylic acid was found in the livers and inguinal adipocyte TAGs of the rats fed SL, whereas it was not present in the rats fed a physical mixture. This suggests that the positional distribution of C8:0 is important in the metabolism of TAGs and may lead to different physiological influences. Conflicting results regarding the influence positional distribution has on absorption and metabolism have been reported, but may be due to differences in the oils and experimental conditions used (Jensen and others 1994; Christensen and others 1995; Tso and others 1999). Straarup and Hoy (2000) extended this examination by comparing the lymphatic transport of a specific SL, comprised of rapeseed oil and decanoic acid (C10:0), a randomized fat, and a physical mixture, in normal and malabsorbing rats. All 3 lipids compared in this study contained the same fatty acid profile. Rapeseed oil was used as the control. In this study, the decanoic acid was located mainly in the sn-1 and sn-3 positions in the specific SL, but at all positions in the randomized lipid, which was synthesized chemically. In normal rats, faster absorption rates of 18:1n-9 and 18:2n-6 fatty acids were obtained from the specific SL and the physical mixture compared to the randomized and control oils. The authors concluded that this was a result of faster hydrolysis of the specific TAG compared to the randomized and native oils. The MCFA and the sn-2 MAG released after hydrolysis of MCTs in the physical mixture may improve the emulsification of LCTs and, therefore, account for the increased absorption rate of the physical mixture. The recoveries of C18:3n-3 were similar for the physical mixture and randomized oils, but significantly lower than for the specific oil. This indicates that the TAG structure of the oil and the presence of C10:0, not the level of individual fatty acids, were the major determinants of the amount of fatty acid absorbed and transported. The specific SL had low amounts of C10:0 at the sn-2 position, and dietary C10:0 from the sn-1 and Vol. 3, 2002 COMPREHENSIVE REVIEWS IN FOOD SCIENCE AND FOOD SAFETY 113

5 CRFSFS: Structured lipids. Comprehensive.. Reviews in Food Science and Food Safety Table 1 Triacylglycerol structure of structured and comparison lipids used in recent studies on enteral and parenteral nutrition. Source Structured Lipid a Comparison Lipid(s) Lee and others :0 or LCFA LCFA 18:2n-6 (63%), 18:1n-9 (23%), 18:3n-3 (6%) 18:2n-6 (66%), 18:1n-9 (24%), 18:3n-3 (5%) 8:0 or LCFA LCFA Soybean Oil Straaup and Hoy :0 10:0 18:1n-9 (42%), 18:2n-6 (22%), 18:3n-3 (10%) 10:0 10:0 10:0 Specific SL LCFA, 10:0 LCFA 10:0 (45%), 18:1n-9 (30%), 18:2n-6 (12%) 18:1n-9 (47%), 18:2n-6 (35%), 18:3n-3 (15%) LCFA, 10:0 LCFA Randomized SL Rapeseed Oil Mu and Hoy :0, 10:0, or 12:0 LCFA 18:1n-9 or 18:2 18:2 (75%), 18:1n-9 (11%) 8:0, 10:0, or 12:0 LCFA Safflower Oil LCFA 18:1n-9 (90%), 18:2 (7%) LCFA High Oleic Sunflower Oil Sandstrom and others 1993, MCFA, LCFA LCFA Bellantone and others 1999, MCFA, LCFA LCFA Chambrier and others 1999, MCFA, LCFA LCFA Kruimel and others 2000, Structolipid (Pharmacia & Upjohn AB, MCFA Rubin and others 2000 Stockholm, Sweden), or FE MCFA (Kabi Pharmacia AB, Stockholm, Sweden) MCFA Intralipid (pharmacia & Upjohn AB) Lipofundin (Braun Melsungen AG, Melsungen, Germany), or Medialipide 20% (B. Braun Inc., Boulongne, France) Lee and others 1999 PUFA LCFA 8:0 (64.3%), 20:5n-3 (17.8%), 22:6n-3 (15%) 18:2n-6 (66%), 18:1n-9 (25 %), 18:3n-3 (5%) PUFA LCFA Soybean Oil Kenler and others 1996, MCFA, 20:5n-3, 20:6n-3 8:0, 10:0 Swails and others 1997 MCFA, 20:5n-3, 20:6n-3 8:0, 10:0 MCFA, 20:5n-3, 20:6n-3 8:0, 10:0 Corn or soybean LCFA Corn or soybean LCFA Corn or soybean LCFA Osmolite HN (Ross Products Division, Columbus, Ohio, U.S.A.) a MCFA = medium chain fatty acid, LCFA = long chain fatty acid sn-3 positions was thought to be absorbed directly through the portal vein; therefore, less C10:0 was expected in the lymph as a result of SL feeding, compared to randomized oil and physical mixture diets. However, similar amounts of C10:0 were observed in the mesenteric lymph from the 3 oils, which indicates better hydrolysis and higher absorption rates of the specific SL, as well as acyl migration in the TAG during hydrolysis. The intragastric administration of fat to the rats excluded the activation of lingual lipase in this study, and rats have only trace amounts of gastric lipase. Therefore, all hydrolysis of test oils measured in this study was a result of pancreatic lipase. The preduodenal lipases preferentially hydrolyze short and medium chain fatty acids from the sn-3 position, so improved recovery from the specific SL would be expected in patients when these lipases are also contributing to hydrolysis. Clinical treatment of short-bowel patients is one promising area for this type of SL, because these patients frequently have compromised EFA status (Jeppesen and others 1997, 1998). Mu and Hoy (2000) compared the intestinal absorption of different SLs in vivo. They studied the lymphatic transport of fatty acids from specific SL containing different MCFA varying from caprylic acid (8:0) to lauric acid (12:0) in the sn-1,3 positions and LCFAs in the sn-2 position, to investigate the effect of chain length of MCFAs on the absorption of LCFAs and the distribution of MC- FAs between the portal vein and lymphatics in rats with normal fat absorption. The results of this study showed that the chain length of MCFAs located in the primary positions does not affect the lymphatic transport of LCFAs in the sn-2 position. This result suggests that similar type SLs may be used to provide different LCFAs according to clinical demand. Similar intestinal absorption of different LCFAs can be expected in other SLs that contain MC- FAs at the sn-1 and sn-3 positions. It appears that it is possible to 114 COMPREHENSIVE REVIEWS IN FOOD SCIENCE AND FOOD SAFETY Vol. 3, 2002

6 manufacture SLs with different MCFAs to direct the fatty acids from the sn-1 and sn-3 positions toward the liver through the portal vein or toward the muscle or adipose tissues through formation of chylomicrons, without affecting the transport of the long-chain fatty acid in the sn-2 position. The TAGs in total parenteral nutrition (TPN) are normally administered as an emulsion. These emulsions are suspected of suppressing the immune function because pneumonia and wound infection often occurs in patients treated with TPN. However, the influence of various lipid emulsions on leukocyte function is still unclear. Kruimel and others (2000) attempted to explain this phenomenon through in vitro studies of lipid emulsions containing fatty acids with different chain lengths on the production of radicals by polymorphonuclear leukocytes. Emulsions made from LCT, a physical mixture of LCT and MCT, and an SL were compared in this study. The results indicated that physical mixtures caused higher peak levels and faster production of oxygen radicals, compared to LCT and SL emulsions. Chambrier and others (1999) conducted a similar study comparing the effect of physical mixtures and SL on postoperative patients. They did not see the hepatic function disturbances in patients given the SL, which are often observed with TPN. The plasma TAG levels remained normal in patients given SL, whereas they significantly increased with the physical mixture. Bellantone and others (1999) gave lipid emulsions to patients after colorectal surgery. The differences between the SL and physical mixture groups were less marked in this study. Structured lipids synthesized from fish oil and MCTs were administered to patients undergoing surgery for upper gastrointestinal malignancies. This diet was compared to a control diet that differed only in its fat source. The SL diet was tolerated significantly better, led to improved hepatic and renal function, and reduced the number of infections per patient (Kenler and others 1996). Lee and others (1999) fed female mice diets supplemented with an SL containing n-3 PUFAs and caprylic acid or soybean oil for 21 d. The effect of the diet on serum lipids and glucose concentrations were determined at the end of the feeding period. In spite of the higher content of caprylic acid in the SL, 8:0 was not detected in the livers of the SL fed mice. High amounts of n-3 PU- FAs were found in the livers of the SL fed mice. These findings suggest that the 8:0 was metabolized quickly for energy, and that different fatty acids (FAs) in the diet may eventually lead to change in fatty acid composition of the liver. SLs containing MC- FAs and n-3 PUFAs could be a therapeutic or medical lipid source, and may be useful in enteral and parenteral nutrition. This SL could decrease serum cholesterols and TAGs. It may also reduce the rate of body weight gain, because the MCFAs were metabolized more rapidly in the body compared to soybean oil. Swails and others (1997) demonstrated significant reductions in prostaglandin production in postsurgical patients receiving an enteral SL formula containing EPA and DHA. This downregulation in prostaglandin production did not predispose the patients to any postoperative events, such as higher incidence of infection or an inability to heal wounds. In fact, improved liver function was reported in the SL patients. The physiological function of the SL appears to be due in part to the n-3 fatty acids acting through altered prostaglandin release from mononuclear cells. Diets containing an SL composed of MCFAs and linoleic acid led to improved absorption of EFAs in patients with cystic fibrosis. These patients experience malabsorption as a result of impaired pancreatic function. However, the SL provided an efficient way to supplement linoleic acid and to provide energy from the MCTs, which were the preferred substrate for oxidative metabolism (Bell and others 1997). No signs of central nervous system toxicity were noted in patients given the SL, and there was no tendency to ketosis (Sandstrom and others 1993). Recently, a long-term study (4 wk treatment) was conducted on the efficacy and safety of an SL prepared from coconut oil and soybean oil. Patients receiving the SL were compared to other patients given LCT. This double-blind, randomized, crossover study indicated that the SL was safe and efficient when provided to patients on home parental nutrition on a long-term basis and that it may be associated with possible reduction in liver dysfunction (Rubin and others 2000). Reduced calorie fats With increasing consumer awareness of the risks associated with high fat intake, a market for reduced calorie fats and fat replacers has opened up. Carbohydrate and protein-based fat replacers are currently available, but cannot be exposed to high temperatures. Therefore, lipid-based fat substitutes are the only option for use in cooking and deep-frying applications and for mimicking all the attributes of a natural fat. Reduced calorie SLs are designed by taking advantage of either limited absorption of long-chain saturates or the low caloric value of SCFAs. The majority of reduced calorie fats and fat substitutes available today contain fatty acids that are not naturally present in edible oils and fats, but may match the chemistry and functions of natural fats. Typically, such products lack nutritionally important EFAs. For example, Akoh and Yee (1997) interesterified tristearin with tricaprin (C10:0) or tricaprylin (C8:0) with sn-1,3-specific immobilized lipase to produce a low calorie SL. One group of researchers (Kanjilal and others 1999) synthesized an SL from natural vegetable oils so it would contain EFAs and natural antioxidants. They incorporated behenic acid into the sn-1 and sn-3 positions of sunflower oil through a transesterification reaction catalyzed by lipases. The synthesized product delivered 5.36 kcal/g and had an improved plastic nature, which increases the potential food applications for such a product, especially since it is a trans-free solid fat. After producing the SL, it was fed to rats and compared to a control group fed sunflower oil. No differences were observed in the amount of food consumed, which indicates that the palatability and taste of the SL was very similar to the native sunflower oil. Additionally, no differences were observed between the groups regarding the levels of major fatty acids of the plasma total lipids. Functional Structured Lipids Plastic fats for food applications Margarine, modified butters, and shortenings are plastic, that is, they have the appearance of solids in that they resist small stresses, but yield to a deforming stress above a certain minimal value (the yield stress) to flow like liquid. The relative proportion of solid to liquid crystals is the dominant controlling factor for hardness, followed by crystal size and polymorphic form (Johnson 1998). Manufacturers need fats with a steep solid fat content (SFC) curve for margarine production. They want their product to be solid in the refrigerator, but spread easily upon removal and melt quickly in the mouth. The spreadability of margarine at refrigerator temperatures is related to its content of solid fats at 2 and 10 C. The solid content at 25 C influences plasticity at room temperature (Brekke 1980). Desired spreadability occurs within a range of roughly 15 to 35% solids (de Man 1992). The fat should crystallize as a polymorph. Interesterification is useful for producing plastic fats and oils suitable for use in margarines and shortenings, because chemical properties of the original fat are relatively unaffected and the fatty acids inherent properties are not changed. Additionally, unsaturation levels stay constant in the fatty acids, and there is no cis-trans isomerization. When short or medium chain fatty acids and LCFAs are incorporated, they can produce TAGs with good Vol. 3, 2002 COMPREHENSIVE REVIEWS IN FOOD SCIENCE AND FOOD SAFETY 115

7 CRFSFS: Structured lipids. Comprehensive.. Reviews in Food Science and Food Safety Table 2 Plastic fats produced by chemical and enzymatic reactions or genetic engineering Structured Lipid Production Method Food Application Reference Stearic Acid and High-Laurate Genetic engineering of high- Production of trans-free Fomuso and Akoh 2001 Canola Oil laurate canola; acidolysis with stearic margarine. acid and a nonspecific lipase High Palmitic Soybean Oil Genetic engineering using Production of trans-free Stoltzfus and others 2000 the fap5 allele margarine-type fats Palm Stearin-Sunflower Oil Transesterification catalyzed by Softening and improved melting Lai and others 1998 sn-1,3 specific and nonspecific characteristics of a component lipases for shortening, pastry, margarine, and other edible fats Stearic Acid and Triolein Acidolysis with sn-1,3 specific Production of trans-free Seriburi and Akoh 1998 lipase; interesterification with stearic margarine-type fats acid methyl ester and a nonspecific lipase Butterfat-Canola Oil Chemical interesterification Improved cold-temperature Rousseau and others 1996 spreadability of butter Triolein and Tripalmitin Enzymatic interesterification in a Adapt technology for Marangoni and others 1993 canola lecithin-hexane modification of vegetable reverse micelle system oils and dairy fat spreadability and temperature stability. Table 2 summarizes experiments conducted using chemical and enzymatic reactions, or genetic engineering to produce plastic fats. Cocoa butter alternatives Cocoa butter is the fat of choice in the confectionery industry. Its polymorphism greatly affects the physical properties of chocolate products, such as gloss, snap, contraction, heat resistance, quick and sharp melting in the mouth, and bloom-resistance (Koyano and others 1990; Loisel and others 1998). The availability of cocoa butter, which affects cost, has prompted much research on alternatives that can be used as cocoa butter replacements or extenders in chocolate and confectionery coatings. There are no naturally occurring fats with similar physical properties to cocoa butter; all alternatives are made by blending and/or modifying fats. When using enzymatic processes for producing cocoa butter alternatives, several factors need to be taken into consideration. The melting behavior has to be very similar to cocoa butter in order to achieve the same cooling effect in the mouth. An alternative fat that is to be used in conjunction with cocoa butter should not interfere with the correct crystallization of the cocoa butter during tempering. crystals are the desirable polymorph in the confectionery industry. The most common cocoa butter equivalents to date include palm oil, palm mid-fractions, illipe (Shorea stenoptera) fat, shea (Butyrospermum parkii) butter, sal (Shorea robusta) fat, and kokum (Garcinia indica) butter. There are also some commercially blended alternatives available. When these natural fat sources are modified by incorporating either palmitic or stearic acid, using sn-1 and sn-3 selective lipases, it is possible to produce a cocoa butter-like fat, in which the fatty acid composition closely resembles that of cocoa butter. An extensive review of cocoa butter alternatives was published by Lipp and Anklam (1998). Foglia and others 1993 suggested beef tallow as a possible base fat for producing SLs that could be used as cocoa butter alternatives. Unpublished data (Osborn and Akoh 2002) from our laboratory indicate that SLs enzymatically made by randomizing beef tallow or by incorporating stearic acid into beef tallow may be useful as a cocoa butter extender, because chocolates produced with the SL had some physical properties similar to chocolates produced with only cocoa butter. Frying oils Kubow (1993) noted that PUFAs and plant sterols may form potentially toxic products when exposed to oxidative stress, either thermally or through aeration. High oleic acid sunflower oil, discussed previously in the genetically engineered SL section, shows excellent behavior with respect to thermooxidation and frying stability (Dobarganes and others 1993). Canola plants that express 80% 18:1 in their seed TAG have much improved heat stability over traditional canola oil. When measured according to the active oxygen method (AOM) for fat stability, conventional soybean oil is rated 10 to 20 h, whereas the high-oleic transgenic lines have been rated up to 140 h (Fitch 1997). Warner and Mounts (1993) found that genetically modified soybean and canola oils had higher flavor characteristics when used in potato frying traditional oils. Other studies have also been published on genetically modified frying oils, such as high-oleic corn oil and low-linolenic soybean oils with improved frying characteristics (Mounts and others 1994; Warner and Knowlton 1997). Physical properties In the above sections on the food applications of SLs, it should have become obvious that methods for testing the physical properties of novel lipids are needed during the development stages. Fats and oils are usually modified to attain a certain functionality, such as improved spreadability, a specific melting point, or a particular solid fat content and temperature profile. It is important to know the thermal characteristics, rheology, crystal habit, texture, and appearance of a new SL when determining its suitability for use in a certain food application. The Mettler dropping point is a simple, yet effective method of measuring the effect of interesterification on fats (AOCS 1989a). In this procedure, liquefied fats are crystallized in sample cups and heated until they reach the dropping point, which is the point when a sample begins to flow under its own weight (Rousseau and Marongoni 1998a). The slip melting point measures the temperature at which a column of fat moves in an open capillary when heated. The drop or slip point of a fat usually occurs at a lower temperature than the melting point, because they will begin flowing at temperatures where 5% solid fat still exists (Timms 1985). Some researchers used dropping point as a verification of complete interesterification (List and others 1995). However, this may not be an accurate measurement for all SLs. Rousseau and others (1996) found that a linear increase in the proportion of canola oil did not lead to a linear reduction in dropping point. The amount of solids in a TAG sample can be determined by pulsed nuclear magnetic resonance (NMR) (AOCS 1989b). Differ- 116 COMPREHENSIVE REVIEWS IN FOOD SCIENCE AND FOOD SAFETY Vol. 3, 2002

8 ential scanning calorimetry (DSC) and dilatormetry techniques are also available to estimate the percentage of solid fats in an SL sample. Knowing the solid fat content of an SL is important for predicting the organoleptic properties it will impart to a food. In order to avoid this problem, fats should be completely melted at 36 ºC (body temperature). The SFC of a fat also contributes to the cooling sensations that will be associated with a product in the mouth. This coldness is important in margarines and chocolate products. In margarines, a large difference between the SFC at 15 and 25 C is correlated to increased cooling sensations (Idris and others 1996). In chocolates, a sharp melting point is indicative of intense cooling sensations in the mouth. As previously mentioned, the polymorphic behavior of an SL is important in food systems. TAGs can form 1 of 3 major polymorphic forms:,, and. Different polymorphs are needed for different food applications. For example, the polymorph should be avoided in margarines, because it causes graininess in the final product (de Man 1983). However, this same crystal is the one desired for chocolate production. Similarly, the crystal, which is associated with softening and low melting points in chocolate products, is the desired polymorph for margarine production. X- ray diffraction is a simple, rapid, and nondestructive method for determining polymorphic structure in fat samples. Although this technique is used for research purposes (Ali and others 2001), the instrumentation is costly and therefore not widely used by industry. DSC is the method of choice for determining crystal structure by most lipid manufacturers. The melting profiles of different cocoa butter crystal polymorphs have been correlated to definitive x-ray determinations (Talbot 1995). Polarized light microscopy can also be used to observe the morphology of crystals in a fat. Interesterification leads to noticeable modifications in crystal morphology. Lard consists of large crystals before interesterification, but contains tiny, delicate crystals afterward (Rousseau and Marangoni 1998a). Viscous and elastic components are measured to determine the rheology of a fat. Rheology plays an important role in the perception of taste, flavor, smoothness, and graininess of margarines, salad dressings, mayonnaise, and chocolate products produced with SLs. Therefore, controlled stress rheology measurements provide a means for relating changes in molecular structure to process behavior and for predicting product performance (Narine and Marangoni 1999). For example, Fomuso and others (2001) recently compared the rheological properties of mayonnaise and Italian salad dressing made from SL (caprylic acid incorporated into olive oil by acidolysis reaction) to the same foods made with unmodified olive oil. The structured lipid mayonnaises and salad dressings displayed similar viscoelastic properties to their olive oil-based counterparts, which indicate similar structural and textural properties between the samples. Fomuso and Akoh (2001) also used rheology to measure the liquid and solid behaviors of trans-free margarine prepared with an enzymatically synthesized SL comprised of high-laurate canola oil and stearic acid. Texture of an SL can be measured with cone penetrometry, an Instron testing machine, or a texture analyzer. Hardness and cohesiveness are important textural measurements for spreadable materials and chocolate products. If an SL is brittle and not spreadable, it will have low cohesive values. Lohman and Hartel (1994) demonstrated a link between SFC of a fat and hardness in chocolate. The same may be true for SL used in chocolate. Mechanical measurements are important for determination of suitability of SLs for certain food applications, but the importance of consumer acceptability cannot be stressed enough. Sensory tests are needed to determine threshold levels for SL incorporation and to describe sensory characteristics such as flavor, odor, and texture. Instrumental analysis of samples may detect no change in physical properties, but the sensory properties of a food could be altered when an SL is substituted for a natural oil. A good example of this phenomenon was documented by Rousseau and Marangoni (1998b). They interesterified canola oil and butterfat to reduce the hardness of butter, but the product led to a rancid, metallic flavor. Free fatty acid removal and steam injection under vacuum successfully removed these undesirable flavors. Unfortunately, all the desirable flavor components were removed at the same time, and the final product was tasteless. The same authors later compared the physical and sensory properties of butterfat SLs produced chemically and enzymatically. Chemical interesterification led to greater changes in physical and sensory attributes than enzymatic interesterification. In both cases, butter flavor degradation must be minimized before the process would be suitable for commercial production of butterfat SLs (Rousseau and Marangoni 1999). Some cocoa butter substitutes have a SFC that decreases over a broader range of temperatures, compared to the sharp decrease in solid fat in cocoa butter near body temperature. Sensory analysis is therefore necessary to determine if such SLs are suitable for use in chocolate manufacturing, because consumers may perceive this remaining solid fat as a waxy mouthfeel. Commercial Outlook for Structured Lipids Commercial product examples Many SLs, produced from a variety of methods, are now commercially available. Genetic engineering allowed Calgene (Monsanto Co.) to market Laurical, which was described in detail in a previous section of this review. Esterification reactions catalyzed by chemicals have also been used to produce SLs commercially. Structolipid (Pharmacia & Upjohn AB, Stockholm, Sweden), comprised of randomized MCFAs and LCFAs, is produced for medical use in TPN treatments. Kabi Pharmacia AB (Stockholm, Sweden) produces a similar product, FE These products provide physicians with alternative lipid sources for TPN formulas, which are normally formulated with LCTs or physical mixtures of MCTs and LCTs. Procter & Gamble s (Cincinnati, Ohio, U.S.A.) Caprenin is a structured lipid containing C8:0, C10:0, and C22:0 fatty acids esterified randomly to a glycerol backbone. Coconut, palm kernel, and rapeseed oils were chemically transesterified to produce this SL, which contains only 5 kcal/g. The behenic acid is partially absorbed by the body and is responsible for the reduced caloric value of the product. Nabisco Foods Group (East Hanover, N.J., U.S.A.) used similar principles to produce Salatrim from C2:0, C3:0, C4:0, and C18:0 fatty acids and glycerol. This product, now marketed as Benefat by Cultor Food Science (Ardsley, N.Y., U.S.A.), provides 5 kcal/g and is intended for use in baking chips, chocolate-flavored coatings, baked and dairy products, dressings, dips, and sauces, or as a cocoa butter substitute (Bell and others 1997). Captex 810D is a reduced calorie SL produced by Abitec Corp. (Columbus, Ohio, U.S.A.). Akoh and others (1998) found that a Captex diet resulted in increased heat production and altered energy metabolism in obese Zucker rats. McKenna and others (1985) also reported improved absorption of 18:2n-6 when Captex was administered to cystic fibrosis patients. Another reduced-calorie SL containing oleic (C18:0) and behenic (C22:0) acids, Bohenin, was designed to slow or prevent fat bloom formation in chocolate products. It promotes the formation of stable crystals and provides 5 kcal/g. The Stepan Food Co. (Maywood, N.J., U.S.A.) markets the Neobee MCTs. The 1095 line prevents bloom formation in chocolate, while the M-5 option prevents uncoated nutritional bars from sticking to packaging material by migrating to the surface of Vol. 3, 2002 COMPREHENSIVE REVIEWS IN FOOD SCIENCE AND FOOD SAFETY 117