Vegetable protein sources for carnivorous fish : potential and challenges.

Abstract

195 Vegetable protein sources for carnivorous fish: potential and challenges S. Refstie and T. Storebakken Institute of Aquaculture Research, 6600 Sunndals�ra, Norway stale.refstie@akvaforsk.nlh.no Summary In this review we describe how carnivorous fishes respond to vegetable protein feedstuffs. The amino acid profiles of gluten and soy products are complementary with respect to amino acid profile, and hold promise for further development of processing to increase protein contents and improve nutritional qualities. Soybeans are rich in antinutritional factors that disturb the digestion and/or physiology of carnivorous fish. Among these are heat_stable factor(s) inducing enteritis. Most soy antinutritional factors are removed by the thermal treatments followed by ethanol washing used to produce protein concentrates. Gluten products contain few antinutritional factors. As with most vegetable protein feedstuffs, gluten and soy protein concentrates do, however, contain phytic acid. Phytic acid_bound phosphorus is unavailable to fish, and phytic acid also binds essential divalent mineral elements, rendering them unavailable. The reduced availability of minerals has possible deleterious consequences and phytic acid should be eliminated by enzymic hydrolysis before feeding diets with high levels of plant protein to carnivorous fishes. Keywords: carnivore, fish, feedstuffs, vegetable protein, antinutritional factors, processing 1993; Kaushik et al. 1995; Olli et al. 1995; Robaina et al . 1995; Burel et al. 1998, 2000a), and low_ glucosinolate rapeseeds (Hardy and Sullivan 1983; Burel et al. 2000a) have been used with success as partial substitutes for fish meal. Feeds for economically important cultured carnivorous fishes are, however, energy dense, and typically contain more than 40% protein, and at least 15% lipid. Because these fish digest starch and/or assimilate and metabolise glucose poorly (reviewed by Wilson 1994), the starch content of the diet is kept low. This calls for vegetable protein feedstuffs with high contents of protein and low contents of starch, sugars, and indigestible non_starch polysaccharides (NSP). Few seed crops meet these criteria, as legumes and oilseeds typically contain 25 to 40% protein and more than 30% carbohydrates. The seed commodity that comes closest is hulled and defatted (hexane extracted) soybean meal, but even this product is unsatisfactorily low in protein and high in NSP and sugars (Table 1). Thus, the most feasible vegetable protein feedstuffs are the industrially manufactured protein concentrates. Corn gluten is already extensively used, and wheat gluten is being introduced. The soy industry has industrialised several processes that concentrate protein (reviewed by Lusas and Riaz 1995), and ethanol washed soy protein concentrates produced from hulled and defatted soy are of high nutritional quality for carnivorous fishes (Storebakken et al. 1998a; Berge et al. 1999; Mambrini et al. 1999; Kissil et al. 2000). Crude and refined soy products are reviewed in detail by Storebakken et al. (2000b). Extraction procedures to concentrate rapeseed protein are also developed (Jones 1979) but, although the resulting rapeseed concentrate is of high nutritional value (Higgs et al. 1994), this process is still not economically viable. Hulling and grinding followed by an air separation process may to some extent concentrate protein from legume seeds with low oil content (Booth et al. 2001). Even when concentrated, the use of plant protein in feeds for carnivorous fish introduces several challenges. Compared with fish meal, the amino acid Introduction Low temperature and steam dried fish meals are the commonly used protein sources in feeds for carnivorous fish. They are of high nutritional value, but are high_ priced, and the supply is limited. Consequently there are major efforts to define and develop cost_effective protein sources that can, at least in part, be substitutes for fish meals in least_cost feed formulations. Problems related to bovine spongiform encephalopathy (BSE) have led restrictions on the use of animal by products in fish feeds. Thus, the main focus is on vegetable protein. Legume seeds, such as peas, lupins, and soybeans (Hughes 1991; Watanabe et al. 1992; Shimeno et al. Recent Advances in Animal Nutrition in Australia, Volume 13 (2001) 196 Refstie, S. and Storebakken, T. composition of vegetable protein is unbalanced, and excessive heating during industrial drying may reduce the protein quality of vegetable feedstuffs even further. More troublesome still is the fact that plants contain various antinutritional factors that disturb the digestion and/or physiology of fish. Carnivorous fishes have short intestines with little microflora (reviewed by Buddington et al. 1997), and therefore are very sensitive to such factors. Protein quality Compared to animal protein, plant protein is generally low in the indispensable amino acids arginine, lysine and methionine. For carnivorous fishes, the reported requirements range from 3 to 6% of crude protein (CP; N x 6.25) for arginine, 4 to 5% of CP for lysine, and 2 to 4% of CP for methionine (National Research Council 1993). As shown in Table 2, these requirements are well met by fish meal, but not by soy protein nor by gluten. The amino acid profiles of corn gluten and soy proteins are, however, complementary. Being a legume, soybean contains high levels of arginine and lysine in its protein, but little methionine. Corn gluten, on the other hand, is deficient in arginine and lysine, but has high methionine content. When combined, these vegetable protein feedstuffs produce a reasonably good partial substitute for fish meal (Watanabe et al. 1993; Akiyama et al. 1995; Yamamoto et al. 1995). Most commercial vegetable protein meals are residues from industrial manufacture of vegetable oil or starch. Such industrial processes usually include repeated heating to soften the seeds, remove solvents, and/or dry the meal residue. Excessive heating reduces the protein quality by amino acid cross_linking, binding to other nutrients, and/or oxidation (Finley and Phillips 1988; Davidek et al. 1990). This destroys the amino acids involved, and reduces the general digestibility of the protein (Lj�kjel et al. 2000). The most vulnerable amino acids are lysine (Maillard reaction) and cysteine (formation of disulfide bonds). When selecting industrially manufactured feedstuffs, it is important to be aware of and consider potential heat damage of the protein. Heat damage is avoided or reduced by moderate, short and moist heating. Vegetable antinutritional factors Vegetable antinutritional factors may broadly be divided into heat_stable and heat_labile factors. Unless these components can be inactivated or removed, they constitute the major restriction on the use of vegetable protein in fish feeds. Table 1 Typical composition of fish meal and vegetable protein ingredients currently used in commercial feeds for car nivorous salmonid fish (percent of dry matter). Protein 78 42 57 68 67 85 2 3 Protein source Fish meal1,2 Full fat soy3,4 Hulled and defatted soy3,4 Soy protein concentrate Corn gluten 2,4 3,4 Oil 12 21 1 1 2 6 4 Starch _ 3 3 7 21 7 5 NSP _ 18 23 19 3 _ Sugars _ 11 14 2 1 _ Wheat gluten5 1 Anderson et al. 1993; Anderson et al. 1992; Lusas and Riaz 1995; Bach_Knudsen 1997: Storebakken et al. 2000a Table 2 Typical critical indispensable (essential) amino acids in protein ingredients currently used in commercial feeds for salmonid fishes as compared to fish meal (g/16 g N = percent of CP). Fish meal1 Arginine Lysine Methionine 1 Soybean 7.3 6.1 1.4 3 1 Corn gluten 3.7 1.8 2.3 2 Wheat gluten 3.6 1.5 1.6 3 5.9 8.1 3.0 2 Lj�kjel et al. 2000; Anderson et al. 1992; Storebakken et al. 2000a Vegetable protein sources for carnivorous fish 197 As the term implies, heat_labile factors may be destroyed or inactivated by thermal processing. Glutens are practically free from such components, but legumes, and particularly soybeans, are rich in heat_labile antinutritional factors. The most significant of these are proteinase inhibitors and agglutinating lectins. Proteinase inhibitors are proteins capable of binding protein_hydrolysing digestive enzymes, thus restricting digestion and utilisation of dietary protein. Lectins are glycoproteins that bind (agglutinate) to receptors in the epithelium of the fish intestine (Hendriks et al. 1990), possibly with deleterious effects. Proteinase inhibitor_ activity is commonly measured as mg bovine trypsin inhibited per g sample (TI_activity; Hammerstrand et al . 1981). Carnivorous fishes are sensitive to proteinase inhibitors (Krogdahl and Holm 1983; Takii et al. 1998), and both nutrient digestibility and growth by carnivorous fish are severely reduced if the dietary TI_activity exceeds 5 mg/g (Wilson and Poe 1985; Krogdahl et al. 1994; Olli et al. 1994). TI_activity may exceed 30 mg/g in raw soybeans. Certain beans, like navy or kidney beans, may also have a high TI_activity, whereas it is negligible in lupins and starch_rich legumes like peas and lentils (Table 3). The most significant heat_stabile antinutritional factors in current vegetable protein feedstuffs for fish are antigens and phytic acid. Indigestible carbohydrates may also be troublemakers, and in particular the NSP. Antigenicity is so far only investigated in salmonid fish, and only for soy products and wheat gluten. Soybeans contain antigenic factor(s) that induce enteritis in the distal intestine of salmonid fishes. This inflammation is characterised by widening and shortening of the mucosal foldings, loss of the supranuclear vacuolisation of the absorptive cells in the intestinal epithelium, widening of and increased amounts of connective tissue in the central stroma within the mucosal foldings, and infiltration of a mixed leukocyte population in the lamina propria and submucosa (Baeverfjord and Krogdahl 1996). It is, furthermore, reflected in elevated lysozyme and IgM levels in the mucosa (Krogdahl et al. 2000), an increased number of circulating leukocyte cells, elevated circulating concentrations of plasma proteins and immunoglobulins, and increased activity of circulating neutrophils, monocytes, and macrophages (Rumsey et al. 1994). The condition is associated with a reduced absorptive capacity for nutrients by the distal intestine (Nordrum et al. 2000). How much the condition actually contributes to the reduced absorption of nutrients seen in soy_fed salmonids (see Figure 1) is unclear, as the distal intestine is not recognised as a major absorptive site in fish (Buddington et al. 1997). Furthermore, salmonid fish suffering the condition appear to grow normally (Refstie et al. 2000, 2001). On the other hand, the large intestine of fish absorbs macromolecules throughout life, and has important enteric immune functions (Buddington et al. 1997). The antigen(s) inducing this inflammatory response are still not identified. It is also unclear whether other plant seeds have similar antigenic properties, and whether the antigens(s) affect fish in general. The antigen(s) are apparently soluble in alcohol, as alcohol washed soy protein concentrates do not induce enteritis, whereas Table 3 Contents in raw and processed legumes of Kunitz tr ypsin inhibitor (TI), Bowman_Birk combined tr ypsin and chymotr ypsin inhibitor (BB_TI), and agglutinating lectins, together with activities of functional TIs, lectins, and urease. Contents of Activities of Lectins mg/g TIs mg/g* 1.3_1.8 1.6 2 Protein source Kunitz_TI mg/g BB_TI mg/g Lectins mg/g** Urease pH rise*** Raw pea 1 1 Raw lentil Raw sweet lupin Raw navy 0.1 11.5 3 bean3 Raw kidney bean Raw soybean 5.4 Conventional1,4,5,6,7 Kunitz inhibitor_free4,5 Lectin_free4 Toasted soybean meal4,6,7,8,9 Soy protein concentrate 6,8 30.3 0.04 28.4 10.7 11.4 13.0 8.3 8.0 <0.0002 17_31 5.6 2.3 2.1 2.0 3_9 2_7 0.0.2 <0.2 * mg bovine tr ypsin inhibited per g meal ** mg lectins agglutinating to br ush border vesicles of chicken per g meal *** pH rise in phosphate buffer as urease acts upon urea to produce ammonia 1 Pisulewska and Pisulewski 2000; 2Booth et al. 2001; 3Dhurandhar and Chang 1990; 4Han et al. 1991; 5Douglas et al. 1999; 6 Anderson and Wolf 1995; 7Maenz et al. 1999; 8Refstie et al. 1999; 9Waldroup et al. 1985 198 Refstie, S. and Storebakken, T. Nitrogen digestibility % of control diet 100 95 90 85 80 y = -0.1475x + 101.88 R2 = 0.84 Lipid digestibility % of control diet 100 95 90 85 80 y = -0.3529x + 100.79 R2 = 0.74 Energy digestibility % of control diet 100 95 90 85 80 0 10 20 30 40 50 % defatted SBM in the diet Figure 1 Reduced digestibility of nitrogen, lipid and energy by Atlantic salmon when replacing low temperature fish meal (LT_FM) by defatted and toasted soybean meal (SBM) in the diet. Fish oil was the only lipid source in all diets. Responses are presented relative to when feeding the LT_FM (control) diet. Data are taken from Refstie et al. 1998, 1999, 2000, 2001, and Storebakken et al. 1998b. y = -0.3258x + 101.76 R2 = 0.91 the alcohol extract (soy velasse) does (Ingh et al. 1996; Krogdahl et al. 2000). Wheat gluten, which may induce celiac sprue in humans, does not induce enteritis in salmon (Storebakken et al. 2000a). Indigestible seed carbohydrates (Table 1) are largely in the form of soluble and insoluble NSP. Dietary soluble NSP may reduce lipid absorption by fish. For instance, indigestible starch (Grisdale_Helland and Helland 1997) and viscous guar gum (Storebakken 1985) restrict the absorption of protein and lipid by salmonids. Impaired lipid absorption when feeding soluble NSP is also a well_known and serious problem in chicken nutrition. Legumes and crucifers furthermore contain indigestible sugars (_galactosides; Table 1) that produce diarrhoea in fish. As shown in Table 4, the phosphorus content of fish meal is substantial. Thus, the concentration in diets for carnivorous fishes is high, and usually exceeds 10 g P/kg DM, but a major fraction of this is present in bone and thus is poorly available (Nordrum et al. 1997; Vielma et al. 1999). This results in a high P load to the aquatic environment, which may lead to algal blooms in freshwater systems (Vielma et al. 2000). Vegetable protein ingredients contain significantly less P than fish meal, and substitution of fish meal by vegetable feedstuffs thus enables the use of less, and more highly digestible, P sources in the diet. As shown in Table 5 this is, however, complicated by the presence of phytic acid which is abundant in all types of vegetable protein feedstuffs that have a potential for use in feeds for carnivorous fish. Phytic acid_bound P is unavailable to fish. Phytic acid also binds essential divalent mineral elements like Zn, and reduces the intestinal hydrolysis and thus the utilisation of dietary protein (Spinelli et al. 1983; Caldwell 1992). Lowered availability of Zn cause eye cataracts, as demonstrated in Chinook salmon (Oncorhynchus tshawytscha; Richardson et al. 1985). Baeverfjord et al. (1998) demonstrated that P deficiency in Atlantic salmon (Salmo salar) is initially manifested as a reduction in whole body Ca and P levels and the development of abnormally soft bones, although growth is little affected. In later stages, growth is severely impaired, and mortality increases. Reduced body Zn concentration is overcome by dietary supplementation of this element (Ramseyer et al. 1999); however, as indicated by Table 5, reduced body concentrations of Ca and P, and thus bone ash, seem harder to overcome. It follows that vegetable antinutritional factors and nutrient classes from different feed ingredients may Vegetable protein sources for carnivorous fish 199 Table 4 Typical phosphorus (P) and phytic acid in dr y matter of fish meal and vegetable protein meals. Ingredient Phosphorus, g/kg Phytic acid, g/kg Phytic acid P as percent of total P 1 2 Fish meal 21.1 _ 1 Soybean2 4.5 7.5 Pea3 4.4 9.7 Rapeseed 14.9 41.5 3 Wheat gluten 2.5 2.1 1 _ 47 3 49 63 24 Storebakken et al. 2000a; Refstie et al. 1999; Burel et al. 2000b Table 5 Availability of nitrogen and phosphorus in and elemental composition of Atlantic salmon grown for 84 days on diets with 60% fish meal or 15% fish meal and 48% untreated or phytase_treated soy protein concentrate as dietar y protein sources. Soy protein concentrate diets were supplemented with dicalcium phosphate to obtain similar dietar y P contents. From Storebakken et al. (1998a). Soy protein concentrate diet 9.3 Phytase treated soy protein concentrate diet 0.5 Dietary phytic acid, g/kg DM* Nitrogen Digestibility, % Retention, % Excretion, g/kg gain Faecal Metabolic Phosphorus Digestibility, % Retention, % Excretion, g/kg gain Faecal Metabolic Whole body concentration of Ash, g/kg Ca, g/kg P, g/kg Mg, mg/kg Zn, mg/kg Ratio Ca : P a,b,c Fish meal diet 0.8 91a 58 ab 85 55 b a b b 88 62 a a ab a 4.8 17.0 41b 30 a 8.2 16.3 c b 5.9 13.8 a a b b 30 25 a a 49 32 a b 8.1 1.4 8.7 0.5 6.3 1.9 b a 21.0 4.2 4.4 403 a a a a 19.2 3.5 4.0 389 b b c b 21.5 4.0 4.4 413 a a b a 61.6 a a 42.6 b b 60.6 a a 0.94 0.85 0.91 Within rows, values not followed by the same superscript are significantly different (P< 0.05) * From dietary wheat interact to affect the overall absorption of nutrients by fish. Although the contribution of each effect is small, the interactional sum may be significant. This is demonstrated in Figure 1, which shows how the digestibility of nutrients by Atlantic salmon is reduced by inclusion of toasted soybean meal in diets based on low temperature fish meal and fish oil. This appears to be a general response in carnivorous fishes, and inclusion of more than 20% soybean meal in the diet usually leads to reduced feed efficiency and slower growth (Watanabe et al. 1992; Shimeno et al. 1993; Kaushik et al. 1995; Olli et al. 1995). From Table 3, it is apparent that raw soybean is the least acceptable protein feedstuff for carnivorous fishes. Defatted soy flakes are, however, typically toasted (steam_cooked) at 105�C for 30 minutes to remove solvent residues after the oil extraction procedure. This reduces the TI_activity down to levels tolerable by carnivorous fishes, and is accompanied by denaturation and inactivation of the lectins. Furthermore, feedstuffs in modern energy_dense fish feeds are subjected to a second moist heating during the high_pressure moist extrusion manufacturing of the diets. Thus, proteinase inhibitors and lectins are rarely 200 Refstie, S. and Storebakken, T. problematic when using defatted soy and/or other vegetable protein feedstuffs in modern fish feeds. Heat_stable antinutritional factors are, however, hard to inactivate, and if not removed they restrict the use of plant proteins in fish feeds. Oligosaccharides, soluble NSP, and soy antigenic factor(s) that induce distal enteritis in salmonids (Ingh et al. 1996) are all soluble in alcohol, and so they are removed by ethanol and water washing in the manufacture of plant protein concentrates. This leaves phytic acid, which should be hydrolysed if vegetable feedstuffs provide a major part of the dietary protein. It is hydrolysed by microbial phytases, which are usually produced from Aspergillus niger. Current phytases are heat labile and consequently susceptible to inactivation during the high_pressure moist extrusion of feeds for carnivorous fish. Furthermore, phytase from A. niger has a temperature optimum of more than 50�C, and its hydrolytic activity is less than 10% of maximum at 10�C (Hoppe 1992). When used as a feed enzyme supplement, the phytase activity will be low at ambient water temperatures where cold_water carnivorous fishes are farmed. An alternative solution is pre_incubation of the vegetable protein sources with phytase, a procedure that is highly effective as shown in Table 5. If culturing fish at warm ambient water temperatures, post_extrusion application of liquid feed enzymes may prove a useful procedure to avoid thermal phytase destruction (Hughes and Soares 1998; Oliva_Teles et al. 1998, Vielma et al. 1998). References Akiyama, T., Unuma, T., Yamamoto, T., Marcouli, P. and Kishi, S. (1995). Combination use of malt protein flour and soybean meal as alternative protein sources of fish meal in fingerling rainbow trout diets. Fisheries Science 61, 828_832. Anderson, R.L. and Wolf, W.R. (1995). Compositional changes in trypsin inhibitors, phytic acid, saponins and isoflavones related to soybean processing. Journal of Nutrition 125, 581S_588S. Anderson, J.S.., Lall, S.P., Anderson, D.M. and Chandrasoma, J. (1992). Apparent and true availability of amino acids from common feed ingredients for Atlantic salmon (Salmo salar) reared in sea water. Aquaculture 108, 111_124. Anderson, J.S., Lall, S.P., Anderson, D.M. and McNiven, M.A. (1993). Evaluation of protein quality in fish meals by chemical and biological assays. Aquaculture 115, 305_325. Bach_Knudsen, K.E. (1997). Carbohydrate and lignin contents of plant material used in animal feeding. Animal Feed Science and Technology 67, 319_338. Baeverfjord, G. and Krogdahl, �. (1996). Development and regression of soybean meal induced enteritis in Atlantic salmon, Salmo salar L., distal intestine: a comparison with the intestines of fasted fish. Journal of Fish Diseases 19, 375_387. Baeverfjord, G., �sg�rd, T. and Shearer, K.D. (1998). Development and detection of phosphorus deficiency in Atlantic salmon, Salmo salar L., parr and post_ smolts. Aquaculture Nutrition 4, 1_11. Berge, G.M., Grisdale_Helland, B. and Helland, S.J. (1999). Soy protein concentrate in diets for Atlantic halibut (Hippoglossus hippoglossus). Aquaculture 178, 139_148. Booth, M.A., Allan, G.L., Frances, J., and Parkinson, S. (2001). Replacement of fish meal in diets for Australian silver perch, Bidyanus bidyanus IV. Effects of dehulling and protein concentration on digestibility of grain legumes. Aquaculture 196, 67_85. Buddington, R.K., Krogdahl, �. and Bakke_McKellep, A.M. (1997). The intestines of carnivorous fish: structure and functions and the relation with diet. Acta Physiologica Scandinavica 161, Suppl. 638, 67_80. Burel C., Boujard, T., Corraze, G., Kaushik, S.J., Boeuf, G., Mol, K.A., Geyten, S.v.d. and K�hn, E.R. (1998). Incorporation of high levels of extruded lupin in diets for rainbow trout (Oncorhynchus mykiss): nutritional value and effect on thyroid status. Aquaculture 163, 325_345. Burel C., Boujard, T., Kaushik, S.J., Boeuf, G., Geyten, S.v.d., Mol, K.A., K�hn, E.R., Quinsac, A., Krouti, M. and Ribaillier, D. (2000a). Potential plant_protein sources as fish meal substitutes in diets for turbot (Psetta maxima): growth, nutrient utilisation and thyroid status. Aquaculture 188, 363_382. Burel, C., Boujard, T., Tulli, F. and Kaushik, S.J. (2000b). Digestibility of extruded peas, extruded lupin, and Future Vegetable protein feedstuffs for carnivorous fishes need to have high protein contents and, in that respect, the most feasible candidates are gluten products. Soy protein concentrates are also promising, and such products are already available commerciallly. As the amino acid compositions of vegetable proteins are generally imbalanced for fish, it is important to combine vegetable protein feedstuffs with complimentary amino acid profiles. This is achieved by combining soy protein and corn gluten. It is also critical to eliminate and remove antinutritional factors. Soybeans are rich in antinutritional factors, but most of these are removed by industrial thermal treatments followed by ethanol washing to produce the commercial protein concentrates. Such manufacturing processes must be optimised to avoid protein damage by excessive heating. It is also desirable to eliminate NSP in vegetable feedstuffs in order to enhance the concentration of digestible energy in the feedstuffs. Finally, phytic acid should be eliminated to avoid disturbances of mineral element absorption and utilisation by fish fed high levels of plant protein. If and when these goals are achieved by cost_effective means, vegetable protein feedstuffs may replace substantial amounts of the fish meal currently used in energy_dense feeds for carnivorous fishes. Vegetable protein sources for carnivorous fish 201 rapeseed meal in rainbow trout (Oncorhynchus mykiss) and turbot (Psetta maxima). Aquaculture 188, 285_ 298. Caldwell, R.A. (1992). Effect of calcium and phytic acid on the activation of trypsinogen and the stability of trypsin. Journal of Agricultural and Food Chemistry 40, 43_47. Davidek, J., Velisek, J. and Pokorny, J. (1990). Chemical Changes during Food Processing. Elsevier, Amsterdam, 448 pp. Dhurandhar, N.V. and Chang, K.C. (1990). Effect of cooking on firmness, trypsin inhibitors, lectins and cysteine/cysteine content of navy and red kidney beans (Phaseolus vulgaris). Journal of Food Science 55, 470_474. Douglas, M.W., Parsons, C.M. and Hymowitz, T. (1999). Nutritional evaluation of lectin_free soybeans for poultry. Poultry Science 78, 91_95. Finley, J.W. and Phillips, R.D. (1988). Protein Quality and Effects of Processing. Marcel Dekker, New York, NY, USA, 137 pp. Grisdale_Helland, B. and Helland, S.J. (1997). Replacement of protein by fat and carbohydrate in diets for Atlantic salmon (Salmo salar) at the end of the freshwater stage. Aquaculture 152, 167_180. Hammerstrand, G.E., Black, L.T. and Glover, J.D. (1981). Trypsin inhibitors in soy products: Modification of the standard analytical procedure. Cereal Chemistry 58, 42_45. Han, Y., Parsons, C.M. and Hymowitz, T. (1991). Nutritional evaluation of soybeans varying in trypsin inhibitor content. Poultry Science 70, 896_906. Hendriks, H.G.C.J.M., Ingh, T.S.G.A.M.v.d., Krogdahl, �., Olli, J. and Koninkx, J.F.J.G. (1990). Binding of soybean agglutinin to small intestinal brush border membranes and brush border membrane enzyme activities in Atlantic salmon (Salmo salar). Aquaculture 91, 163_170. Higgs, D.A., Dosanjh, B.S., Prendergast, A.F., Beames, R.M., Hardy, R.W., Riley, W. and Deacon, G. (1994). Use of rapeseed/canola protein products in finfish diets. In: Nutrition and Utilization Technology in Aquaculture, pp.130_156 (eds. C. Lim, D.J. Seesa). AOAC Press, Champaign, USA. Hoppe, P.P. (1992). Review of the biological effects and the ecological importance of phytase in pigs. Presentations at the 4th BASF Animal Nutrition Forum, November 4_5, 1992. BASF, Ludwigshafen, Germany. Hughes, S.G. (1991). Use of lupin flour as a replacement for full_fat soy in diets for rainbow trout (Oncorhynchus mykiss). Aquaculture 93, 57_662. Hughes, K.P. and Soares, J.H. (1998). Efficacy of phytase on phosphorus utilization in practical diets fed to striped bass Morone saxatilis. Aquaculture Nutrition 4, 133_140. Ingh, T.S.G.A.M.v.d., Olli, J.J. and Krogdahl, �. (1996). Alcohol_soluble components in soybeans cause morphological changes in the distal intestine of Atlantic salmon, Salmo salar L. Journal of Fish Diseases 19, 47_53. Jones, J.D. (1979). Rapeseed protein concentrate preparation and evaluation. Journal of the American Oil Chemists Society 56, 716_721. Kaushik, S.J., Cravedi, J.P., Lalles, J.P., Sumpter, J., Fauconneau, B. and Laroche, M. (1995). Partial or total replacement of fish meal by soybean protein on growth, protein utilization, potential estrogenic or antigenic effects, cholesterolemia and flesh quality in rainbow trout, Oncorhynchus mykiss. Aquaculture 133, 257_ 274. Kissil, G.W., Lupatsch, I., Higgs, D.A. and Hardy, R.W. (2000). Dietary substitution of soy and rapeseed protein concentrates for fish meal, and their effects om growth and nutrient utilization in gilthead seabream Sparus aurata L. Aquaculture Research 31, 595_601. Krogdahl, �. and H.H. Holm (1983). Pancreatic proteinases from man, trout, pig, cow, chicken, mink and fox enzyme activities and inhibition by soybean and lima bean inhibitors. Comparative Biochemistry and Physiology 74B, 4409. Krogdahl, �., Lea, T.B. and Olli, J.J. (1994). Soybean proteinase inhibitors affect intestinal trypsin activities and amino acid digestibilities in rainbow trout (Oncorhynchus mykiss). Comparative Biochemistry and Physiology 107A, 215_219. Krogdahl, �., Bakke_McKellep, A.M., R�ed, K.H. and Baeverfjord, G. (2000). Feeding Atlantic salmon Salmo salar L. soybean products: effects on disease resistance (furunculosis), and lysozyme and IgM levels in the intestinal mucosa. Aquaculture Nutrition 6, 77_84. Lj�kjel, K., Harstad, O.M. and Skrede, A. (2000). Effect of heat treatment of soybean meal and fish meal on amino acid digestibility in mink and dairy cows. Animal Feed Science and Technology 84, 83_95. Lusas, E.W. and Riaz, M.N. (1995). Soy protein products: Processing and use. Journal of Nutrition 125, 573S_ 580S. Maenz, D.D., Irish, G.G. and Classen, H.L. (1999). Carbohydrate_binding and agglutinating lectins in raw and processed soybean meals. Animal Feed Science and Technology 76, 335_343. Mambrini, M., Roem, A.J., Crav�di, J.P., Lall�s, J.P. and Kaushik, S.J. (1999). Effects of replacing fish meal with soy protein concentrate and of DL_methionine supplementation in high_energy, extruded diets on the growth and nutrient utilization of rainbow trout, Oncorhynchus mykiss. Journal of Animal Science 77, 2990_2999. National Research Council (1993). Nutrient Requirements of Fish. National Academy Press, Washington, USA, 114 pp. Nordrum, S., �sg�rd, T., Shearer, K.D. and Arnessen, P. (1997). Availability of phosphorus in fish bone meal and inorganic salts to Atlantic salmon (Salmo salar) as determined by retention. Aquaculture 157, 51_61. Nordrum, S., Bakke_McKellep, A.M., Krogdahl, �. and Buddington, R.K. (2000). Effects of soybean meal and 202 Refstie, S. and Storebakken, T. salinity on intestinal transport of nutrients in Atlantic salmon (Salmo salar L.) and rainbow trout (Oncorhynchus mykiss). Comparative Biochemistry and Physiology 125B, 317_335. Olivia_Teles, A., Periera, J.P., Gouveia, A. and Gomes, E. (1998). Utilisation of diets supplemented with microbial phytase by seabass (Dicentrarchus labrax) juveniles. Aquatic Living Resources 11, 255_259. Olli, J.J., Hjelmeland, K. and Krogdahl, �. (1994). Soybean trypsin inhibitors in diets for Atlantic salmon (Salmo salar, L): effects on nutrient digestibilities and trypsin in pyloric caeca homogenate and intestinal content. Comparative Biochemistry and Physiology 109A, 923_928. Olli, J.J., Krogdahl, �. and V�ben�, A. (1995). Dehulled solvent_extracted soybean meal as a protein source in diets for Atlantic salmon, Salmo salar L. Aquaculture Research 26, 167_174. Pisulewska, E. and Pisulewski, P.M. (2000). Trypsin inhibitor activity of legume seeds (peas, chickling vetch, lentils, and soya beans) as affected by the technique of harvest. Animal Feed Science and Technology 86, 261_265. Ramseyer, L., Garling, D., Hill, G. and Link, J. (1999). Effect of dietary zinc supplementation and phytase pre_ treatment of soybean meal or corn gluten meal on growth, zinc status and zinc_related metabolism in rainbow trout (Oncorhynchus mykiss). Fish Physiology and Biochemistry 20, 251_261. Refstie, S., Storebakken, T. and Roem, A.J. (1998). Feed consumption and conversion in Atlantic salmon (Salmo salar) fed diets with fish meal, extracted soybean meal or soybean meal with reduced content of oligosaccharides, trypsin inhibitors, lectins and soya antigens. Aquaculture 162, 3312. Refstie, S., Svihus, B., Shearer, K.D. and Storebakken, T. (1999). Non_starch polysaccharides in soybean meals and effects on the absorption of nutrients in farmed Atlantic salmon and broiler