Expanding the utilization of sustainable plant products in ... · compounded rate of 9.2% per year...

Expanding the utilization of sustainable plant products

in aquafeeds: a review

Delbert M Gatlin III1, Frederic T Barrows2, Paul Brown3, Konrad Dabrowski4,T Gibson Gaylord1,RonaldW Hardy5, Eliot Herman6, Gongshe Hu7, —shild Krogdahl8, Richard Nelson9,Kenneth Overturf1, Michael Rust10,Wendy Sealey5, Denise Skonberg11, Edward J Souza7,David Stone5, RichWilson9 & EveWurtele12

1Texas A&M University System, College Station,TX, USA2USDA/ARS Hagerman Fish Culture Experiment Station, Hagerman, ID, USA3Purdue University,West Lafayette, IN, USA4The Ohio State University, Columbus, OH, USA5Hagerman Fish Culture Experiment Station, Hagerman, ID, USA6USDA/ARS ^ Donald Danforth Plant Science Center, St Louis, MO, USA7USDA/ARS,Washington, DC, USA8Aquaculture Protein Centre, —s, Norway9Nelson & Sons, Murray, UT, USA10NOAA Northwest Fisheries Science Center, Seattle,WA, USA11University of Maine, Orono, ME, USA12Iowa State University, Ames, IA, USA

Correspondence: D M Gatlin III, Department of Wildlife and Fisheries Sciences, 2258 TAMUS, College Station, TX 77843-2258, USA.

E-mail: [email protected]

Gatlin and Barrows are Chair andVice-chair, respectively, of the Plant Products in AquafeedsWorking Group, and coordinated the devel-

opment of this document; all other authors are listed in alphabetical order.

Abstract

Continued growth and intensi¢cation of aquacultureproduction depends upon the development of sus-tainable protein sources to replace ¢sh meal inaquafeeds. This document reviews various plantfeedstu¡s, which currently are or potentially may beincorporated into aquafeeds to support the sustain-able production of various ¢sh species in aquacul-ture. The plant feedstu¡s considered includeoilseeds, legumes and cereal grains, which tradition-ally have beenused as protein or energyconcentratesas well as novel products developed through variousprocessing technologies.The nutritional compositionof these various feedstu¡s are considered along withthe presence of any bioactive compounds that maypositively or negatively a¡ect the target organism. Li-pid composition of these feedstu¡s is not speci¢callyconsidered although it is recognized that incorporat-ing lipid supplements in aquafeeds to achieve properfatty acid pro¢les to meet the metabolic requirementsof ¢sh and maximize human health bene¢ts are im-portant aspects. Speci¢c strategies and techniques to

optimize the nutritional composition of plant feed-stu¡s and limit potentiallyadverse e¡ects of bioactivecompounds are also described. Such informationwillprovide a foundation for developing strategic re-search plans for increasing the use of plant feedstu¡sin aquaculture to reduce dependence of animalfeedstu¡s and thereby enhance the sustainability ofaquaculture.

Keywords: alternative proteins, sustainable aqua-feeds, plant proteins

Introduction/justification for increaseduse of plant products in aquaculture

According to the UN Food and Agriculture Organiza-tion, aquaculture is growing more rapidly than allother animal food-production sectors (FAO. State ofworld aquaculture 2006). Its contribution to globalsupplies of ¢sh, crustaceans and molluscs increasedfrom 3.9% of total production by weight in 1970to 33% in 2005. This growth is at an average

Aquaculture Research, 2007, 38, 551^579 doi:10.1111/j.1365-2109.2007.01704.x

r 2007 TheAuthorsJournal Compilationr 2007 Blackwell Publishing Ltd 551

compounded rate of 9.2% per year since 1970, com-pared with only 1.4% for capture ¢sheries and 2.8%for terrestrial farmed-meat production systems. It isremarkable that one out of every three ¢sh consumedin the world is now farm raised.The expansion of aquaculture production has

been accompanied by rapid growth of aquafeedproduction. The challenge facing the aquaculture in-dustry is to identify economically viable and environ-mentally friendlyalternatives to ¢sh meal and ¢sh oilon which many present aquafeeds are largely based.While the supply of ¢sh meal and oil is arguably sus-tainable, the anticipated growth in demand interna-tionally for use in aquafeeds is expected to exceed thesupply in the next decade.Thus, the aquafeeds indus-try has recognized for many years that viable utiliza-tion of plant feedstu¡s formulated in aquafeeds forthe production of cold, cool and warmwater aquaticspecies is an essential requirement for future devel-opment of aquaculture. Such plant feedstu¡s mustprovide nutritious diets that will e¡ectively growaquatic species with minimal environmental impactand produce high-quality ¢sh £esh to confer humanhealth bene¢ts in a cost-e¡ective manner. As theaquaculture industry continues to expand on a glo-bal scale, access to key feedstu¡s, such as ¢sh mealand oil, will become increasingly limited because ofa ¢nite wild-harvest resource. In addition to con-cerns about the sustainability of ¢sheries resources,other issues including the potential presence of or-ganic and inorganic contaminants in ¢sh meal andthe net e¡ect of demand-and-supply economics inthe global market require enhanced e¡orts to thor-oughly evaluate reasonable alternatives such as var-ious plant feedstu¡s.The US soybean industry has identi¢ed the poten-

tial demand for plant protein in aquafeeds as an op-portunity to increase the use of US soybeans. Since1995, the United Soybean Board (USB) has fundedmarket-development activities, primarily in China.This programme has increased demand for soybeanmeal (SBM) for farm-raised ¢sh from almost 0 to anestimated 5 million metric tonnes (mmt) in 2005. In2002, USB created a Managed Aquaculture Pro-gramme, the Soy-in Aquaculture-Initiative (SIA), withthe joint objectives of expanding its marketing initia-tive and conducting research on the factors in SBMthat limit its use in salmonids. Nevertheless, there arestill unidenti¢ed biologically active compounds thatlimit the amount of SBM and other plant feedstu¡s indiets of carnivorous species, including most marinespecies of commercial importance (existing or poten-

tial) as well as salmonids. For reasons that are notclear, carnivorous species have limited tolerance toSBM that ranges from none in juvenile chinook sal-mon (Oncorhynchus tshawytschaWalbaum) to possiblysomewhat higher in diets for rainbow trout, O.mykissWalbaum (Olli & Krogdahl 1994), and Atlantic sal-mon, Salmo salar L. (Olli, Krogdahl, van dan Ingh &Bratt�s1994; B�verfjord & Krogdahl 1996; Krogdahl,Bakke-McKellep & Baeverfjord 2003). The salmonaquaculture industry is clamoring for a stable sourceof protein in terms of cost and supply, and such asupply would bene¢t other established and emergingsectors of the aquaculture industry as well.

Targeted nutritional and non-nutritionalcharacteristics (including cost) of plant-derived feedstuffs for inclusion inaquafeeds

Fish meal has been the protein source of choice inaquafeeds for many reasons, including its high pro-tein content, excellent amino acid pro¢le, high nutri-ent digestibility, general lack of antinutrients, relativelow price (until recently) and its wide availability.Plant-derived feedstu¡s all have some characteristicsthat place them at a disadvantage to ¢sh meal interms of their suitability for use in aquafeeds. How-ever, demand for protein ingredients is expected toexceed the annual world supply of ¢sh meal in thenext decade, and this increased demand will changethe economic and nutritional paradigms that up tonow have resulted in high use levels of ¢sh meal inaquafeeds. The plant feedstu¡s considered in thisdocument include oilseeds, legumes and cerealgrains, which traditionally have been used as proteinor energy concentrates as well as novel products de-veloped through various processing technologies. Li-pid composition of these feedstu¡s is not speci¢callyconsidered although it is recognized that incorporat-ing lipid supplements in aquafeeds to achieve properfatty acid pro¢les to meet the metabolic requirementsof ¢sh and maximize human health bene¢ts areimportant aspects.To be a viable alternative feedstu¡ to ¢sh meal

in aquafeeds, a candidate ingredient must possesscertain characteristics, including wide availability,competitive price, plus ease of handling, shipping,storage and use in feed production. Furthermore, itmust possess certain nutritional characteristics,such as low levels of ¢bre, starch, especially non-soluble carbohydrates and antinutrients, plus have a

Utilization of plant products in aquafeeds DMGatlin III et al. Aquaculture Research, 2007, 38, 551^579

r 2007 TheAuthors552 Journal Compilationr 2007 Blackwell Publishing Ltd, Aquaculture Research, 38, 551^579

relativelyhigh protein content, favourable aminoacidpro¢le, high nutrient digestibility and reasonable pa-latability. Although some plant-derived ingredients,such as soy protein concentrate (SPC) or wheat glu-ten, possess most of these characteristics, they havebeen too expensive relative to the price of ¢sh mealto be used in most aquafeeds. It is likely that a combi-nation of plant-derived feed ingredients will be re-quired to replace ¢sh meal, and that supplements,such as amino acids, £avourings and possibly exo-genous enzymes, will be needed to produce aqua-feeds without ¢sh meal that support growth ratesnecessary for the economic production of farmed¢sh. Table 1 summarizes the nutrients found in ¢shmeal (expressed on an as-fed basis), and the rangeof nutrient concentrations that alternative ingredi-ents should have to be viable alternatives to ¢shmeal.Various grain and oilseed crops are grown in theUnited States, and their recent production is sum-marized in Table 2. The gross nutrient compositionof these and other plant feedstu¡s are presented inTable 3.

Candidate plant products for increaseduse in aquafeeds

Soybean

Soybean, Glycinemax Linnaeus, is the leading oilseedcrop produced globally and its projected production

for 2004^2005 is expected to exceed 200mmt. Alarge part of this production is used in the extractionof oil yielding a cake of high protein quality. This isprocessed to yield a wide array of soybean products,such as soy £our, SBM, SPC and soy protein isolate(SPI) that have been evaluated in ¢sh. Full-fat SBM(heat-treated whole soybeans) also has been evalu-ated in ¢sh. Soybean meal has been the predominantform of soybean used and is available either as de-hulled (� 48% crude protein) or with hulls added(� 44% crude protein) (NRC1993).Soy products are regarded as economical and

nutritious feedstu¡s with high crude protein contentand a reasonably balanced amino acid pro¢le. How-ever, certain nutritional characteristics and thepresence of several antinutritional factors merit dis-cussion. Concentrations of the 10 essential aminoacids (EAA) and tyrosine are generally lower in SBMthan in ¢shmealwith the exceptionof cystine, whichis present at higher concentrations in SBM. The EAAof concern are lysine, methionine and threonine thatmay be limiting in soy-based diets fed to aquatic ani-mals. Concentrations of these EAA increase withprocessing of soy £akes to SPC and SPI and approachor exceed those found in ¢sh meal. However, due tothe processing costs involved, these products are notyet economical for large-scale use in aquafeeds.Crude fat and ash concentrations of solvent-extracted SBM and other soy products tend to belower than those in ¢sh meal, but carbohydrateconcentrations tend to be higher. Lower ash and fat

Table 1 Proximate and nutrient content (as-fed basis) of¢sh meal, and targeted ranges in alternative ingredientsderived from grains and oilseeds

Category/nutrient

Fishmeal(menhaden)

Target rangeforalternativeingredients

Crude protein (%) 65–72 48–80

Crude lipid (%) 5–8 2–20

Fibre (%) o2 o6

Ash (%) 7–15 4–8

NFE (%) o1 o20

Starch (%) o1 o20

Non-soluble CHO (%) None o8

Arginine (%) 3.75 43.0

Lysine (%) 4.72 43.5

Methionine (%) 1.75 41.5

Threonine (%) 2.5 42.2

o-3 fatty acids � 2% �

�The fatty acids provided by alternative protein feedstu¡s varyconsiderably and lipid supplements will be the primary sourceof fatty acids.

Table 2 Grain and oilseed production in the United States

Grain oroilseed

Acresplanted

Production(�1000 bushels)�

Metrictonnes

Barley 3 875 000 352 445 7 689 709

Canolaw

Corn 81 759 000 9 761 085 248 463 980

Cottonseed 14 195 400 7 711 980

Oat 4 246 000 114 878 1 667 450

Peanut

Peas/lupinsw 1 657 000 4 821 250 2 186 880

Rice 3 384 000 223 235 10 125 770

Soybeans 72 375 000 2 756 794 75 185 291

Sunflower 2 709 000 4 018 355 1 822 700

Wheat 65 871 000 2 550 383 69 555 900

�A bushel of barley, corn, soybeans and wheat weighs 48, 56, 60and 60 pounds respectively.wProduction of canola, peas and lupins in the United States isnegligible.Source: USDA Agricultural Statistics (2005), http://www.nass.usda.gov.

Aquaculture Research, 2007, 38, 551^579 Utilization of plant products in aquafeeds DMGatlin III et al.

r 2007 TheAuthorsJournal Compilationr 2007 Blackwell Publishing Ltd, Aquaculture Research, 38, 551^579 553

concentrations in soy products can be overcome byappropriate supplementation with mineral premixesand lipids, but high concentrations of carbohydratesremain an area of concern. Carbohydrates in soy-beans are largely present as oligosaccharides suchas sucrose, ra⁄nose and stachyose. Sucrose is gener-ally available to aquatic animals, but ra⁄nose andstachyose are not digestible due to a lack of a-galacto-sidases that are necessary to metabolize these com-plex sugars. An added concern is the low availabilityof phosphorus and cationic minerals that are largelyunavailable due to their being bound in or by phyticacid.The addition of the enzyme phytase to feeds hasbeen shown to improve phosphorus and cationicmineral availability. There are di¡erences in concen-trations of vitamins between soy products and ¢shmeal.Vitamin availability data in ¢sh are scarce andfeeds are generally forti¢ed with vitamins assumingminimal availability from feedstu¡s (NRC1993).

Barley

Barley,Hordeumvulgare Linnaeus, is grown for multi-ple purposes including animal feeds, human food,malting for the production of alcoholic beverages orinclusion in extracts and syrups for adding £avour,sweetness and colour to a variety of prepared foods.Barley is a short-season, early-maturing crop grownon both irrigated and dry land production areas inthe United States, primarily in the Northern Plainsstates and in the Paci¢c Northwest. The United Statesis the seventh-largest barley-producing country intheworld, but is in the top ¢ve of exporting countries.

Feed barley is included in rations for many animalspecies, but has not been widely used in aquafeeds.Several factors have limited its use, but improve-ments in crop genetics and processing technologyare improving the value of this ingredient for aqua-feeds. New varieties are available including hull-less (lower ¢bre) and low phytic acid varieties(Bregitzer 2005), both of which will help the aquacul-ture industry decrease nutrients and solids in e¥uent.Waxy type varieties also are available in which

the ratio of amylose to amylopectin has been altered.This may have a positive e¡ect on energy availability,but also a¡ects the ease of manufacturing. The waxybarley provides considerable binding capabilitiesfor feeds produced by cooking extrusion. This isbene¢cial for the production of high protein andenergy diets, because less carbohydrate is needed forpellet production (Barrows & Hardy 2001).While barley contains a good nutrient pro¢le in its

native state, the co-products resulting from the devel-opment of other productswill allowsuchbarleyco-pro-ducts to be included in higher amounts in aquafeeds.For example, although corn has been the traditionalsource of feed-stock for ethanol production, barleyalsoshows considerable promise (Hicks, Taylor, Kohout,Kurantz,Thomas, O Brien, Johnston, Flores, Moreau &Hoot 2004). In the Northern Plains and Paci¢c North-west, barley may be the most economical source offermentable carbohydrates for ethanol production.Methods are beingdeveloped to concentrate the proteinfrombarley before fermentation for ethanol production(Flores, Eustace & Hicks 2004). Owing to the aminoacid pattern and protein content, this protein concen-trate will be of highvalue to the aquaculture industry.

Table 3 Typical composition (as-fed basis) of ¢sh meal and various plant feedstu¡s

IngredientInternationalfeed number

Drymatter(%)

Protein(%)

Lipid(%)

Ash(%)

Lysine(%)

Methionine(%)

Cystine(%)

Fish meal, herring� 5-02-000 92.0 72.0 8.4 10.4 5.57 2.08 0.74

Barleyw 4-00-552 88.0 14.9 2.1 2.9 0.44 0.16 0.24

Canola� 5-06-145 93.0 38.0 3.8 6.8 2.27 0.70 0.47

Corn� 4-02-935 88.0 8.5 3.6 1.3 0.25 0.17 0.22

Corn gluten meal� 5-28-242 91.0 60.4 1.8 2.1 1.11 1.63 1.20

Cottonseed meal� 5-01-619 92.0 41.7 1.8 6.4 1.89 0.50 0.45

Lupin Lupinus angustifoliusz (whole) 5-27-717 89.0 39.2 10.3 2.8 1.40 0.27 0.51

Field peasz (whole) 5-03-600 89.0 25.6 1.3 3.4 1.50 0.21 0.31

Soybean meal, de-hulled� 5-04-612 90.0 48.5 0.9 5.8 3.08 0.68 0.75

Soy protein concentratew 90.0 64.0 3.0 1.5 4.20 0.90 1.00

Wheat� 4-05-268 88.0 12.9 1.7 1.6 0.36 0.21 0.27

�Data from NRC (1993).wData from NRC (1998).zData from Allan et al. (2000).



Barley also contains high levels of b-glucans,which have been shown to have strong healthbene¢ts (Lupton, Robinson & Morin 1994; Jenkins,Kendall, Vuksan, Vidgen, Parker, Faulkner, Mehling,Garsetti,Testolin, Cunnane, Ryan & Corey 2002). Bar-ley can be consumed directly to obtain these bene¢ts,but b-glucans are being extracted, puri¢ed and man-ufactured into pills for the nutritional-supplementindustry. This process, like ethanol production, willresult in a protein concentrate as a co-product withhigh nutritional value.

Canola

Canola, Brassica rapa Linnaeus, is produced from cul-tivars of rapeseed that have been bred to contain lowlevels of erucic acid and glucosinolates, a group ofantinutrients that interfere with normal thyroidfunction. Canola seed is an oilseed, and canola oil isthe primary product of its cultivation. After oil is ex-tracted and the solvent removed, the resulting meal(10% moisture) contains about 3.5% residual oil,35% crude protein, 6% ash and 12% crude ¢bre onan ‘as-is’ basis. It also contains about 4% phytic acidand15 mmol of glucosinolates g�1. The price of cano-la meal is linked to that of SBM, and re£ects the pro-tein content of the meal. Canola meal is not widelyused in aquafeeds in the United States as it is notwidely available, but is used in Canadian aquafeeds.Canola meal can be further processed to produce a

protein isolate, called canola protein concentrate(CPC). Canola protein concentrate has been widelytested as a protein source for salmonids and othercarnivorous species of farmed ¢sh. As the primaryprotein source in aquafeeds for salmon and trout,CPC supports growth rates similar to those of ¢shfed ¢sh meal-based diets, as long as amino acid sup-plements are used to overcome limiting amino acidlevels and providing feeding stimulants, such as be-taine, are added to the diet to overcome reduced in-take. Canola protein concentrate has a proteincontent similar to high-quality ¢sh meal. At present,CPC is not widely available for use in aquafeeds, andmarket prices have not been established.

Corn

Production of corn, Zea may Linnaeus, is higher thanany other grain or oilseed in the United States, but incontrast to other grains, only a small percentage ofannual production is consumed directly by humans.

Corn oil is the primary food product of corn produc-tion and although most corn grown in the UnitedStates is fed to livestock as an energy source, an everincreasing amount is going to ethanol production.Corn starch is used to produce over 400 products,including ethanol, paper coatings and corn syrup,awidely used sweetener in foods and beverages.In the wet-milling process, the corn kernel is sepa-

rated into its main components, bran/¢bre, germ, glu-ten and starch.The oil is separated from the germ andthe remaining corn germ meal is used as an ingredi-ent or added to corn gluten feed.The gluten protein isconcentrated, ¢ltered and dried to form corn glutenmeal.The commonly traded cornglutenmeal is guar-anteed to contain a minimum of 60% protein on an‘as-is’ basis. Re¢ned and puri¢ed corn gluten proteincan have a crude protein content of 70^73% crudeprotein, but there are limits in the commercial pro-duction process so these levels cannot always be met.A higher crude protein corn gluten meal would be amore suitable product for use in aquafeeds althoughthe price would be higher than commercially tradedcorn gluten meal. Economic and marketing analysisis required to determine if the production of a high-protein corn gluten meal is warranted.Corn gluten meal is currently widely used in aqua-

feeds for salmon and several marine species such asEuropean sea bass, Dicentrarchus labrax Linnaeusand gilthead sea bream, Sparus aurata Linnaeus.Corn gluten meal protein is highly digestible, but de-¢cient in the EAA lysine. This characteristic and itscrude protein content result in upper inclusion limitsof 20^25% of diet for corn gluten meal in aquafeedsfor salmon and marine ¢sh. Generally, use levels arein the10^15% range.Corn distillers dried grains with solubles (DDGS) is

the ending co-product of ethanol production. Con-ventional DDGS contains 28^32% crude protein andis relatively high in ¢bre content, which limits itsuse in aquafeeds. Ethanol production is expected toincrease substantially in the United States, andresearch is underway to explore the possibility ofremoving protein and ¢bre from starch before usingthe starch to produce ethanol. If this approachbecomes widespread, corn protein concentrate maybecome more readily available for use in aquafeeds.

Cottonseed

Cottonseed, Gossypium hirsute Linnaeus, is the thirdleading legume seed by weight (after soybean andrapeseed) used worldwide. Owing to its high protein



value for human consumption (Alford, Liepa & Van-beber1996) and animals, as well as low market pricein comparisonwith other legumes and ¢sh meal, cot-tonseed meal (CSM) consequently has an immensepotential for incorporation in high-protein aqua-feeds. The results of numerous studies evaluatingCSM in cat¢sh, salmonid and tilapia diets indicatethat between 10% and 30% of solvent extracted,40% protein CSM can be used in aquaculture dietswithout growth depression. However, these studiesmost frequently used the strategy of replacing oneplant protein source with the other; whereas, morechallenging and pro¢table would be to replace ¢shmeal, or other animal protein sources. Lee, Dabrows-ki, Blom, Bai and Stromberg (2002) used an approachto entirely replace ¢sh meal with a mixture of animalby-product and both SBM and CSM in diets for juve-nile rainbow trout with extremely positive results.However, the authors also pointed out that the sourceof CSM and related processing in£uences utilizationof CSM in ¢sh diets. Lee, Rinchard, Dabrowski,Babiak, Ottobre and Christensen (2006) summarizeda series of studies in rainbow trout where CSM re-placed ¢sh meal entirely over the 3-year period with-out signi¢cantly impacting growth rate of female andmale rainbow trout, although diets with CSM hadsigni¢cantly lower protein and phosphorus assimila-tion. In one study, channel cat¢sh, Ictalurus punctatusRa¢nesque, were raised at high densities in earthenponds and fed to satiation a diet containing 51%CSM with supplemented lysine (0.65%). Results indi-cated that growth rate, dressout percentage and che-mical composition of the ¢llets did not di¡ersigni¢cantly from ¢sh fed diets containing SBM(42%) (Robinson & Li 1994). Fowler (1980) demon-strated that up to 34% of CSM can be used in the dietfor twoPaci¢c salmon species to replace SBMwithoutgrowth depression.

Peas/lupins

Peas, Pisum sativum Linnaeus, and lupin, Lupinus sp.Linnaeus, are common plant ingredients either al-ready being used or considered for use in aquafeedfor marine and freshwater ¢sh throughout the world(Allan & Rowland 1994; Allan, Parkinson, Booth,Stone, Rowland, Frances & Warner-Smith 2000;Glencross, Boujardand & Kaushick 2003; Glencross,Curnow, Hawkins, Kissil & Peterson 2003; Thiessen,Campbell & Tyler 2003; Dias, Rueda-Jasso, Panserat,Conceicao, Gomes & Dinis 2004; Glencross, Carter,

Duijster, Evans, Dods, McCa¡erty, Hawkins, Maasand& Sipsas 2004; Glencross, Evans, Hawkins & Jones2004). Both products are produced in signi¢cantquantities throughout the world. The 2004 produc-tion levels of peas and lupin were 9.1 and 1.0mmtrespectively (FAOSTAT 2005). US production of peasin 2004 was 0.9mmt, while lupin production wasinsigni¢cant for the same period.The nutrient pro¢le of peas and lupin indicate that

that theyhave the potential to replace signi¢cant pro-portions of ¢sh meal protein in aquafeeds. Whencompared with animal meals such as ¢sh meals,meat and bone meals and poultry meals, the proteincontent of peas and lupin are moderate, lysine andmethionine are limited, and both contain high levelsof carbohydrate (Allan et al.2000). Carbohydrates arethe primary component of peas and lupin (Table 3).Starch is the predominant energy reserve carbohy-drate in peas, and may comprise approximately 55%of the total seed. In contrast, lupins contain insignif-icant levels of starch (o3%), with non-starch poly-saccharides (NSPs) such as b-(1,4)-galactan actingas the primary energy storage carbohydrate (Anon-ymous1992; van Barneveld1999). In general, ¢sh donot have the capacity to metabolize NSPs of plant ori-gin (Stickney & Shumway1974; Kuz’mina1996).Studies investigating the nutritional value of peas

and lupin as ¢sh meal protein replacements for aqua-feeds have been favourable. Thiessen, Campbell andTyler (2003) reported protein apparent digestibilitycoe⁄cient (ADC) values in excess of 90% for a rangeof pea products for rainbow trout, and also demon-strated that dietary levels of 25% dehulled peas couldbe used to replace SBM. Allan et al. (2000) reportedexcellent protein ADC values from lupin (495%) bythe freshwater silver perch, Bidyanus bidyanusMitch-ell. However, Glencross, Curnow et al. (2003) reportedprotein ADC values of495% for a range of lupinvari-eties by the red sea bream, Pagrus auratus Forster.Glencross, Curnow et al. (2003) also reported accepta-ble growth of juvenile red sea breamwhen fed lupinatup to 60% of diet. In all of the above-mentioned stu-dies, energyADCswere low for peaand lupinproducts(range 51^78%). This is indicative of the high level ofindigestible carbohydrates present in these products.

Wheat

Wheat (Triticum aestivum Linnaeus. and T. diccoides,var. durum) is widely produced in the United Statesand unlike other cereals has at least six large



subdivisions of cultivars that are based on end-useproperties of the grain. Current production of breadwheat (T. aestivum) in the United States is approxi-mately 60mmt (NASS http://usda.mannlib.cornel-l.edu/usda/). Although wheat is primarily milled forhuman consumption,7^9% of US production is usedin animal feeds as grain and nearly all of mill feed(the 20^25% of kernel that is a by-product of £ourproduction) is used in animal diets. The functionalcharacteristic of wheat that di¡erentiates it fromother cereal grains is the presence of large seed sto-rage protein that forms the cross-linking networksof gluten when hydrated £our is mixed. Gluten givesdough it elastic characteristics and the ability to re-tain gasses produced by leavening agents. Gluten isalso an e¡ective binder of aquafeeds because of itsstrength and limited water solubility.Traditional pre-parations of wheat for diets have utilized £aking (rollmilling) of whole grains or pelleting of mill feed andgrains. A rapidly increasing use of wheat in Europeand North America is cereal fractionation, theseparation of vital wheat gluten from other seedcomponents and further decomposition into starch,soluble ¢bre and oil. Although still a relatively smallportion of total usage in North America (o1%),usage of components is increasing driven by bakingindustries that are seeking greater control over the¢nished product either for quality control or nutri-tional marketing. The availability of wheat fractionscould increase the £exibilityof formulations for aqua-culture, particularly for feed mills and aquacultureareas proximal to grain fractionation facilities.The wheat kernel in publications of composition is

often assumed to be 12% protein by weight at 12%moisture (FAO 1970; Bushuk & Wrigley 1974). How-ever, grain used in animal diets is often composed ofless-costly soft wheat, which tend to be low in proteinas a class (o10%) or lower protein lots (o11%) ofhard endosperm wheat that failed to make millinggrade. Like other cereals, whole grain wheat is pri-marily an energy source based on its high starchcomposition (typically 470%) rather than a sourceof protein. In addition to protein and starch, wheatkernels are approximately 2% crude ¢bre and o2%ash (Martin, Leonard & Stamp 1976). The greatestcomponents of the ash fraction are potassium(0.40%), phosphorus (0.38%) and sulphur (0.20%),with calcium (0.04^0.05%) as a relatively minormineral constituent (Martin et al. 1976; Guttieri, Bo-wen, Dorsch, Souza & Raboy 2004). Phosphorus inwheat is primarily sequestered in the aluerone layerin phytin, phytic acid with chelated metal ions. Phy-

tic acid mutants in wheat can signi¢cantly reducephytic acid concentration and enhance inorganicphosphorus levels that are nutritionally more readilyavailable in the diet (Guttieri et al. 2004). However,previous work in grain crops (Raboy & Dickinson1993) and more recent studies in wheat and barleysuggest that reducing phosphorus fertilization mayreduce overall phytic acid levels in cereals whileincreasing nutritionally available phosphorus.Protein quality in wheat is similar to other cereals

with glutamic acid accounting for 30% of total pro-tein weight (FAO1970). Nutritionally important ami-no acids, lysine and methionine are relatively minorconstituents of the total, 3% and 1.5% respectively(FAO 1970). The amino acid composition of mill feedis improved over whole grain because the proteinsare primarily enzymatic and structural proteinsderived from the bran and embryo rather than thestorage proteins of the endosperm. Wheat embryoproteins contain as much as 8% lysine and 2%methionine (Pomeranz, Carvajal, Hoseney & Ward1970). Mill feed, however, has several drawbacks inaquafeeds. First, the energy quotient is reduced be-cause most of the digestible carbohydrate is removedin the production of £our. Secondly, the glutenin andgliadin storage proteins that form gluten are also re-moved in themilling process with the remaining pro-teins of the bran generallyof poorer value for use as afeed binder. The remaining mill feed is high in crude¢bre (� 10%,Martin et al.1976) of limited digestibilityfor monogastric animals and has a more concen-trated phytic acid composition that reaches nearly0.5% of total weight (Guttieri et al. 2004).

Bioactive compounds in and/ornutritional limitations of plant products

Soybean

Carbohydrate fractions

Non-starch polysaccharides. Soybeans are charac-terized by a high content of NSPs and negligiblestarch. The high NSPs content represents a majorchallenge for the use of soybean products in ¢shdiets. This fraction provides marginal energy for the¢sh due to limited microbial fermentation, and maynegatively a¡ect nutrient utilization and reduce feede⁄ciency. A review of soy products (Storebakken,Refstie & Ruyter 2000) indicated raw soybeans con-tain in the range of 20% NSPs, approximately one-third of which are soluble and another fraction is



highly viscous. Elevated faecal water content hasbeen reported in salmonids when fed diets withSBM (Olli, Hjelmeland & Krogdahl1994; Refstie, Stor-ebakken & Roem 1998; Storebakken, Kvien, Shearer,Grisdale-Helland, Helland & Berge1998), which prob-ably in part is related to the NSPs content of SBM.Furthermore, Refstie, Svihus, Shearer and Storebak-ken (1999) attributed a trend of reduced digestibilityof fat and protein in both Atlantic salmon and broilerchickens to increased intestinal viscosity due to diet-ary soybean NSPs. Soybean NSPs increased the visc-osity of the intestinal contents of chickens, which isan indicator of, and possibly a major reason for, thedisturbed absorption of nutrients in birds fed dietsrich in soluble NSPs such as mixed-linked b-glucansfrom barley (Almirall, Francesch, Perez-Vendrell,Brufau & Esteve-Garcia 1995) and arabinoxylans inrye (Bedford & Classen 1992) or wheat (Choct,Hughes,Wang, Bedford, Morgan & Annison1996). Insome ¢sh, however, soybean NSPs have been shownto increase intestinal and faecal water contents with-out a¡ecting the intestinal viscosity. In spite of this,the antinutritional e¡ects were the same includingreduced digestibility of fat and nitrogenwith soy pro-ducts containing the most NSPs. Similar but more se-vere e¡ects on nutrient absorption and faecal watercontent occur in rainbow trout when using NSPssuch as guar gum and alginates as binders in aqua-feeds (Davies1985; Storebakken1985; Storebakken &Austreng1987).The antinutritional actions of soluble NSPs are still

not fully understood. Increased viscosity of the in-testinal contents may subsequently impair di¡usionand convective transport of digestive enzymes andsubstrates within the gastrointestinal contents. How-ever, antinutritional e¡ects also probably are asso-ciated with binding or trapping of and subsequentincreased faecal excretion of steroids, in particularbile acids, as seen with galactomannans in rats(Levrat, Moundras, Younes, Morand, Demigne &Re¤ me¤ sy 1996; Favier, Bost, Guittard, Demigne¤ &Re¤ me¤ sy 1997; Moundras, Behr, Re¤ me¤ sy & Demigne¤1997). As NSPs do not appear to a¡ect intestinalviscosity in ¢sh, this latter e¡ect may be responsiblefor the reduced nutrient absorption with NSPs-rich diets. Non-starch polysaccharides also may havesimilar e¡ects as do oligosaccharides on waterbalance in saltwater ¢sh; these e¡ects could be addi-tive when feeding diets with oligosaccharide-con-taining soy meals. Increased drinking has beenindicated by an increase in faecal output of sodiumin Atlantic salmon fed diets with high amounts of

oligosaccharide-free SPC, independent of dietaryphytate concentration (Storebakken et al.1998).

Oligosaccharides. Soybean meal contains oligo-saccharide levels of up to 15% (Russett 2002). Incontrast, oligosaccharide levels in SPC are � 3%(Russett 2002). Stachyose and ra⁄nose are the twomost predominant oligosaccharides in SBM. Oligo-saccharides commonly found in SBM have been re-ported to be indigestible to ¢sh (Refstie et al. 1998).They also have been reported to increase the viscos-ity of the chime in the digestive tract and interferewith the uptake of nutrients, particularly fat andminerals, in salmonid species (Refstie et al. 1998;Refstie, Sahlstrom, Brathen, Baeverfjord & Krogedal2005). The oligosaccharide component of SBM alsohas been linked with reduced growth performance(Refstie et al. 1998) and the occurrence of SBM-induced enteritis in several salmonid ¢sh species(van den Ingh, Krogdahl, Olli, Hendrix & Koninkx1991; van den Ingh, Olli & Krogdahl 1996; Bureau,Harris & Cho1998). In Atlantic salmon, the enteritisdevelops rapidly during the ¢rst weeks of exposureeven at SBM inclusion levels as low as 10% of dietand regresses to apparently normal histology 3weeks after introduction of a diet without SBM (B�-verfjord & Krogdahl 1996).Van den Ingh et al. (1996)fed Atlantic salmon a series of diets containing 26%SBM that had been exposed to a range of heat treat-ments to denature trypsin inhibitor as well as a ¢shmeal-based diet in which the alcohol-soluble compo-nents of SBM were added back, and two control ¢shmeal diets containing no soy products. Control dietsfailed to induce any morphological changes in theintestinal tract whereas morphological changeswere evident in all ¢sh fed diets containing SBM andthe diet containing the alcohol-soluble extract. Thealcohol-soluble extract contained oligosaccharidessaponins and possibly low levels of other antinutri-ents, known as well as unknown.Van den Ingh et al.(1996) suggested that the antinutritional factor wasin the alcohol-soluble component of the SBM andconcluded that either saponins or oligosaccharideswere possible candidates. However, in a trial screen-ing with isolated soybean antinutrients, oligosac-charides, trypsin inhibitors, saponins or lectin didnot show any enteritis-inducing e¡ect (Krogdahl1996).As soybean oligosaccharides are indigestible and

are linked with reduced nutrient uptake and growthperformance as well as with the occurrence ofsoybean-induced enteritis, ultimately it would be



desirable to have a soybean product where theoligosaccharide component has been altered orremoved completely. Fermentation, by eitherbacterial or fungal organisms (Refstie et al. 2005, R.Barrows, pers. comm.), may be used to reduce thenegative e¡ects of soybean oligosaccharides on nutri-ent digestibility and growth performance in ¢sh.Oligosaccharides also may be removed during theproduction of SPC.

Seed proteins including antigenic compounds, proteaseinhibitors and lectins

Often, when SBM inclusion in aquafeed exceeds acertain level, impaired utilization occurs due to re-duced feed intake, growth and development of intest-inal enteritis (Bakke McKellep, Press, Baeverfjord,Krogdahl & Landsverk 2000). These conditions maybe related to protein toxicity by antimetabolicproteins such as lipoxygenase, lectins, urease andtrypsin inhibitors as well as possibly by an immuno-logical-based food intolerance that is essentially afood allergy. Trout and salmon fed diets containinghigh levels of SBM, for example, may exhibit skin le-sions, gastrointestinal tract alterations and excessmucus excretion in the faeces. This mimics the food-allergy response to soybean in some humans, swineand mice where atopic skin response or hives anddiarrhoea with massive mucus is a prompt responseto allergic challenge, with continuing exposure lead-ing to intestinal morphology changes in the form ofenteritis, which may lead to autoimmune disease.Recognition of a food intolerance that is immuno-

logically rather than bioactivity based imposes a dif-ferent set of problems and solutions to utilizing SBMin aquafeeds. Most bioactive proteins are easily inac-tivated by heat treatment that is part of feed produc-tion. However, immunological intolerance may onlyrequire the physical presence of the protein and notits enzymatic activity; therefore, heat and otherprocessing treatments will not signi¢cantly alterallergenicity.

Antigenic proteins. Allergenic responses in mam-mals to seeds (Herman 2004) almost always involvesensitivity tomultiple di¡erent proteins. Seed-proteinintolerance is often not restricted to a single sensitiz-ing crop species but rather it is more oftenmanifestedas a broader intolerance extending across species.The large majority of seed protein content resultsfrom the expression of gene families including the le-gumin and vicilin storage proteins, prolamine sto-

rage proteins, lectins, trypsin inhibitors and oleosin,a protein associated with seed oil storage. The genesthat encode these proteins from diverse species arevery similar in sequence and the resulting proteinproducts similar in three-dimensional structure andsequence. Because of the similarity of the proteins’structure among species, antibodies produced in re-sponse to protein exposure fromone species are oftencross-reactive with the homologous protein of an-other species.Globular storage proteins in soybeans such as gly-

cininand b-conglycinin or their breakdownproductsmay be involved in causing allergic reactions in ani-mals. The severity of such reactions varies consider-ably in various aquatic species. The e¡ects of theseproteins are reduced following heat treatment andprocessing.

Protease inhibitors. Legume seeds, as many otherplant seeds, are equipped with one or more proteaseinhibitors that will inhibit proteolytic enzymes in thegastrointestinal tract of pests that commonly attackthe seeds in ¢elds (Liener 2000). As digestive pro-teases have been highly conserved through evolu-tion, the plant protease inhibitors such as trypsininhibitors also inhibit proteases in the gastrointest-inal tract of monogastric animals. Enzymes of ¢shseem to be particularly sensitive to the proteaseinhibitors (Krogdahl & Holm 1983). Heat treatmentwill inactivate protease inhibitors to a major extent.However, the heat needed for total inactivation ofthe inhibitors may reduce protein quality throughoxidation and fusion with other components of thesoybean. A compromise must therefore be made be-tween inactivation and deterioration of protein qual-ity, which may leave some inhibitors active. It is likelythat the remaining inhibitors in concert with the in-digestible heat-damaged protein are signi¢cant fac-tors for explaining the generally lower nutrientavailability of standard SBM and increased require-ments for sulphur-containing amino acids oftenrecognized in animals fed SBM (Drackley 2000).Although without compromising growth and pro-tein digestibility, low levels of protease inhibitors insalmonid diets have been observed to triggerincreased secretion of pancreatic proteases indicat-ing increased use of metabolic resources, whereas athigher levels both protein digestibility and growth iscompromised (Olli, Krogdahl et al. 1994). Investiga-tion of processes taking place within the soybeanproteins in general and protease inhibitors in parti-cular during heat treatment and mechanisms of



interference between digestive processes and pro-ducts of heat treatment may lead to improved proces-sing techniques for soybeans.

Lectins. One of several antinutritional factorsfound in soybeans and other plant feedstu¡s are thebioactive group of glycoproteins known as lectins(NRC 1993; Francis, Makkar & Becker 2001). Owingto their ability to agglutinate red blood cells they arealso referred to as agglutinins or in the case ofsoybeans, soybeanagglutinins (SBA). Concentrationsof SBA in SBM are reported to be in the range of10^200 mg g�1.Lectins possess the ability to bind reversibly and

speci¢cally to carbohydrates and glycoconjugates,which is responsible for their numerous physiologi-cal e¡ects. For example, lectins bind avidly withintestinal glycoproteins on the epithelial surface andinterfere with nutrient absorption. Lectins easilyavoid digestion and then pass into the intestinewhere they may bind with the epithelium. Somelectins may cause disruption of membrane integrityand the initiation of a cascade of immune andautoimmune events that ultimately lead to celldeath. Lectins gain entry into circulation andmay bind to organs. They may also increase perme-ability of membranes to other proteins increasingincidence of allergic reactions. In addition to thedisruption of intestinal membrane integrity,lectins suchas SBA have beenobserved in some casesto cause hypertrophy, hyperplasia, decreased nutri-ent uptake, decreased intestinal enzyme activity,increase gut length and mass as well as increasepancreatic secretions. Lectins also may have ane¡ect on systemic insulin levels and may mimic itse¡ects. In rats, for example, increasing doses of kid-ney bean lectins resulted in depressed plasma levelsof insulin.The e¡ects of SBA on ¢sh have not been as exten-

sively studied as with terrestrial animals. However,early studies demonstrated alteration in intestinalstructure and changes in immune function in ¢shfed high levels (60^70%) of SBM in their diets. Thesediets also brought about decreased intestinal enzymeactivity and decreased plasma insulin levels. How-ever, these studies did not indicate the speci¢c causeof the observed responses. In another study, rainbowtrout and Atlantic salmon were fed diets containingSBM (60%) and an SBM-free diet containing SBA ata concentration that would be encountered in anSBM-based diet at the 60% inclusion level. Signi¢-cant damage to intestinal membranes was observed

for both dietary treatments in both species of ¢sh,indicating the involvement of SBA in the observedresponse.In more recent studies, rainbow trout and Atlantic

salmon were fed puri¢ed diets containing a range ofSBA concentrations that would be found in SBM at35% and 40% inclusion rates respectively. There wasno evidence of damage to the intestinal epithelium ineither species. Serum insulin levels in both fastedand fed ¢sh were not a¡ected by dietary lectin levelsin rainbow trout, but a decline in serum insulinlevels with increasing dietary lectin levels wasobserved in Atlantic salmon. Results from variousstudies suggest that damage to intestinal structuremay be occurring at SBM inclusion levels that are440% of diet.

Lectins, as discussed earlier, are resistant to pro-teolytic enzymes, but are either inactivated or de-stroyed by heat treatment. They are not a¡ected byheating to 60 1C but are destroyed when heated to100 1C for 5min. Lectins in a variety of legumesother than soybeans are not destroyed to any appre-ciable degree by extrusion. This is of great relevanceas a majority of aquafeeds are manufactured usingthe process of extrusion. Another avenue that is openis the utilization of lectin-free SBM in aquafeeds.However, lectin-free lines of soybeans have not yetbeen evaluated in ¢sh and present a future researchopportunity.

Oestrogenic compounds

Oestrogenic activity of soybean in ¢shwas suspectedbased on vitellogenin appearance in male sturgeonfed commercial (animal/plant-based) diets; however,Pelissero, LeMenn and Kaushick (1991) can be cred-ited for the ¢rst direct evidence. Pelissero, Bennetau,Babin, LeMenn and Dunogues (1991) extended those¢ndings by directly examining oestrogenic activity(vitellogenin secretion in juvenile sturgeon) of iso£a-vones and cumestans by intraperitoneal injections ofgenistein, daidzein, equol and coumestrol and com-paring with oestradiol. Oestradiol (0.0001mgg�1

body weight), coumestrol (0.05mgg�1) and genis-tein (0.2mgg�1) injections resulted in similar vitel-logenin increase, approximately 103 compared withthe control (vehicle).When genistein aglycone was added into rainbow

trout diets at 500 or 1000mg kg�1 diet, there wereno di¡erences in blood plasma oestradiol concentra-tions of females throughout a 12-month samplingperiod whereas males showed many di¡erences in



steroid pro¢les (decrease in 11-ketotestosterone andtestosterone) before and at the time of spawning(Bennetau-Pelissero, Breton, Bennetau, Corraze, Le-Menn, Davail-Cuisset, Helou & Kaushik 2001). Iso£a-vonoids also may a¡ect steroidogenesis regulationand the amounts of circulating hormones by in£uen-cing sex steroid-binding proteins in plasma. However,a⁄nity of genistein in competition with oestradiolis four orders of magnitude lower in Arctic charr(Tollefsen, Ovrevik & Stenersen 2004). Human andrainbow trout plasma showed very weak bindinga⁄nities for daidzein and genistein, where concen-trations higher than103 of oestradiol were needed todisplace natural steroids (Milligan, Khan & Nash1998).Oestradiol glucuronides were 10 times less potent

than oestradiol when evaluated in sturgeon hepato-cyte cultures producing vitellogenin (Latonnelle,LeMenn, Kaushik & Bennetau-Pelissero 2002). How-ever, when evaluated as relative a⁄nity to oestrogenreceptors, i.e. displacing 50% of radiolabelled oestra-diol, oestradiol glucuronide was 103 less e⁄caciousin case of rainbow trout hepatic nuclear extract, and1.5 � 103 less e⁄cacious in case of sturgeon pre-parations (Latonnelle, Fostier, LeMenn & Bennetau-Pelissero 2002).Bennetau-Pelissero, Latonnell, Lamonthe, Shin-

karuk-Poix and Kaushik (2004) analysed total,glucuronidase aryl-sulphatase hydrolysed, genisteinand daidzein concentrations in SBM and, based onan ELISA procedure, estimated a total of 2.6 �0.27mgg�1. This is very close to the concentrationof iso£avonoids in SPI, 2.98 � 0.1mgg�1 (Fang,Yu &Badger 2004). The latter authors also estimated thatmore than 95% of genistein and daidzein are presentas glycosides. The evidence in mammals suggeststhat in vivo exposure to iso£avonoids, or their poten-tial e¡ects on target tissues, requires glycoside hydro-lysis in the gut, and deglucuronidation by microbial£ora, and reabsorption before enterohepatic recy-cling to presume signi¢cant physiological role ofthese substances. Only fragmentary information isavailable concerning these processes in ¢sh in orderto understand the potential role in vivo. D’Souza,Skonberg, Camire, Guthrie, Malison and Lima (2005)reported that gensitein aglycone supplement in rain-bow trout diets resulted in genistein accumulation in¢sh muscle although neither dose^response norduration of feeding were evident to have an e¡ect.However, metabolismand tissue deposition of dietarygenistein aglycone as used in that study may di¡erfrom that of genistein glycosides, which are present

in soybeanproducts.Thus, further research is neededto directly compare dietary sources of genisteinglycosides and aglycones with respect to hydrolysis,deposition and hepatic metabolism in ¢sh.Two studies on the e¡ect of genistein on reproduc-

tive performance of striped bass (Pollack, Ottinger,Sullivan & Woods III 2003) and European eel(Tzchori, Degani, Elisha, Eliyahu, Hurvitz, Vaya &Moav 2004) are somewhat controversial. First, no ne-gative e¡ect of genistein aglycone at a dose as high as8000mg kg�1diet was observed in striped bass after6weeks of feeding.Tzchori et al. (2004) used only 2 or20mg gensitein kg�1 diet with juvenile eel of initialweight of 0.5 g expecting an oestrogenic e¡ect (fem-inization) on sex di¡erentiation. The authors claimedthat the lower dose signi¢cantly increased the pro-portion of females (55%) in comparisonwith controls(5%).

Phytic acid

Approximately two-thirds of the total phosphorus inoilseed meals or grains and their by-product meals ispresent as phytic acid (phytate), the main storageform of phosphorus in seeds. The bioavailability ofphytate phosphorus to ¢sh and to virtually all mono-gastric animals is very low, with reports of phos-phorus availability in SBM ranging from virtuallynil (Riche & Brown1996) to 22% available (Sugiura,Dong, Rathbone & Hardy1998). Furthermore, phyticacid lowers the availability of certain divalent ca-tions, notably zinc, to carnivorous species of ¢sh(trout and salmon) and to omnivorous species (cat-¢sh), and also has been reported in some studies toreduce the apparent digestibility of protein. Heattreatment associated with extrusion pelleting doesnot improve the availability of phytate phosphorusin oilseed meals or grains or reduce antagonistic in-teractions with other essential nutrients. The onlye¡ective strategies to reduce the e¡ects of phytic acidin SBM or other seed-derived products is to removephytic acid by processing, hydrolyse it using the en-zyme phytase or to utilize single-gene mutant vari-eties of grains and oilseeds in which a smallerpercentage of total phosphorus in the seed is storedas phytate phosphorus. Low-phytate cultivars ofcorn, barley, wheat and soybeans have been devel-oped but are not widely available as livestock feed in-gredients. Microbial phytase is more e¡ective atincreasing phosphorus bioavailability in soybeanand canola meals, and in wheat middlings, than inwhole grains, such as wheat or barley (Cheng &



Hardy 2002). However, there are issues related to theheat sensitivity of phytase and its temperature ofactivity that must be resolved before phytase isregularly used in aquafeeds.Soybeanmeal contains approximately 4% phytate.

Phytate is concentrated in the process used to pro-duce SPC (and other protein concentrates), becausethe phytate fraction stays with the protein fraction.Thus, SPC contains approximately 7^10% phytate, a100% increase over SBM.

Amino acid limitations

Soybean protein is well known to be limiting in totalsulphur amino acids (TSAA; methionine plus cy-steine) when utilized in animal feeds. Soybean mealand SPC each contain TSAA at approximately 2.95%of protein. These levels have been demonstrated to bede¢cient in awide variety of ¢sh species.Supplemental methionine can overcome part of

the growth reduction observed when a large percen-tage of the total dietary protein is from soy origin. Forexample, Cai and Burtle (1996) demonstrated thatan SBM^corn diet was de¢cient in methionine forchannel cat¢sh and supplementation of DL-methio-nine could improve growth rates and feed e⁄ciency.Mambrini, Roem, Cravedi, Lalles and Kaushik (1999)also demonstrated a positive e¡ect of DL-methioninesupplementation when SPC was the primary proteinsource for rainbow trout. Although ¢sh in neitherof these experiments could attain growth ratesequivalent to control ¢sh receiving some ¢shmealprotein, DL-methionine supplementation did have apositive e¡ect. Keembiyehetty and Gatlin III (1997)demonstrated that hybrid striped bass, Moronechrysops � M. saxatilis, could e¡ectively utilize 55%SBM in the diet along with ¢shmeal and supplemen-tal L- or DL-methionine to meet the TSAA require-ment.Supplementation of multiple amino acids to diets

containing over 42% extruded SBM also has beendemonstrated to alleviate growth retardation inrainbow trout (Yamamoto, Shima, Furuita & Suzuki2002), whileTakagi, Shimeno, Hosokawa and Ukawa(2001) found that growth retardation in red seab-ream could be partially overcome when a high levelof SPC (52%) was included in the diet. These authorsfound that methionine appeared to be the ¢rst limit-ing amino acid for growth. Additional supplementa-tion with lysine also improved growth, indicating itmay be limiting as well. Floreto, Baker and Brown(2000), working with American lobster, Homarus

americana Edwards, also determined that multipleamino acid supplements (methionine, arginine andtryptophan) to diets containing high levels of SBMimproved growth and feed e⁄ciency. Furuya, Pezza-to, Barros, Pezzato, Furuya and Miranda (2004) ob-served similar responses with Nile tilapia whenfeeding more than 60% SBM. These authors supple-mented a ¢sh meal-free diet with methionine, lysine,threonine and dicalcium phosphate to attain growthrates and feed e⁄ciency equal to ¢sh consuming the¢sh meal-based control diet.As noted across multiple species, cysteine can

meet 40^60% of the TSAA requirement (Wilson2002). Other sulphur-containing compounds thatare metabolic derivatives of methionine and cysteinealso may be limiting in aquafeeds when soy proteinsare the primary protein source. Soybeans and otherplant products do not contain taurine, which hasbeen demonstrated to be conditionally indispensablefor some ¢sh species (Takeuchi 2001). Current re-search also has indicated that taurine may be limit-ing in all-plant protein diets, even for rainbow trout,which have some capacity to synthesize taurine fromcysteine.

Saponins

In general, saponins have high toxicity to ¢sh whenapplied externally (Roy, Munshi & Dutta 1990) andpreparations of saponin extracts from tea seed cakehave been used to eradicate predacious ¢sh in ponds(Chen & Chen 1997). The majority of work on sapo-nins has concentrated on Quillaja saponin (derivedfrom the Quillaja saponaria Molina tree) because ofits use in vaccine adjuvants (Oda, Matsuda, Muraka-mi, Katayama, Ohgitani & Yoshikawa 2003). Quillajasaponins, when administered before or with humangamma globulin (HGG), was able to increase absorp-tion and appearance of HGG in the blood plasma ofNile tilapia (Jenkins, Harris & Pulsford1991). Quillajasaponins were recognized for their toxicity; however,supplementation of 0.6% cholesterol into a diet forchicks with 0.6% Quillaja saponins restored feedintake and growth of animals.Bureau et al. (1998) concluded that 1500 and

3000mg of Quillaja saponins kg�1 diet added to sal-monid feed already containing 32% of SPC causedsigni¢cant intestinal damage and higher saponinconcentrations depressed growth of chinook salmonand rainbow trout. These results require site-by-sitecomparison of puri¢ed soyasaponins and Quillajasaponins, to provide conclusive evidence with diets



where no SPC is included, as it may be argued thatSPC saponins had anadditive negative impact on dietacceptance and utilization by salmonids.Twibell and Wilson (2004) followed this ¢nding

with studies on channel cat¢sh and were able to de-crease the growth-depressing e¡ect of SBM (55% ofthe diet) when cholesterol supplement was added tothe formulation. However, there was no e¡ect of2600mg of puri¢ed soyasaponins (97% of soyasapo-nin B) on the feed intake of channel cat¢sh. Francis,Makkar and Becker (2005) examined the role ofdietary Quillaja saponin addition on performance,growth and fecundity of tilapia and common carp.The comparison of all dietary treatments was basedon ¢ve individual ¢sh per diet housed in metabolicchambers. It was apparent that control ¢sh-depressed growthwas directly associated with frequ-ent mouth-brooding episodes during experiment,whereas in other treatments with less females, ¢shdid not spawn any eggs. Therefore it can be con-cluded that growth data were not representative forgroup-feeding tilapia and confounded by an unevensex ratio in treatments and reproduction of females.Therefore, at this time it can be concluded that thereis no substantial, scienti¢cally sound evidence tosuggest any positive role of Quillaja saponins in ¢shnutrition.

Barley

The inclusion of barley in aquafeeds has been limitedby several factors. Low protein, high ¢bre, b-glucanand phytate contents are important considerationsfor barley. These factors can be accommodated byspeci¢c formulations or through use of improvedcultivars or products.

Nutrient composition

The protein content of barley ranges from 9% to15%depending on variety, and also can vary due to envir-onmental conditions. The higher protein varietieshave increased value in aquafeeds, especially for car-nivorous ¢sh, as protein needs of these species are of-ten in excess of 40% of diet (NRC1993). Lysine is the¢rst limiting amino acid in barley protein at approxi-mately 3.6% of protein. For rainbow trout, this levelwill yield a de¢cit of approximately 0.35% lysine inthe diet, which can be overcome by blending proteinsources with high lysine contents or adding supple-mental lysine. Arginine also will be marginally de¢-

cient in barley protein. The TSAA content of barleyappears adequate for ¢sh but the skewed ratio ofmethionine to cysteine in favour of cysteine may re-quire some supplementation of methionine to thediets of ¢sh utilizing high levels of barley protein(when available). The ¢bre content of barley (6%) istwice that of wheat (NRC 1994). Fibre is indigestibleto coldwater ¢sh and only increases faecal output.De-hulling barely or the use of hulless varieties willreduce the ¢bre content making it more likely to beused in aquafeeds.The value of barley to rainbow trout is increased

when the feed is manufactured by cooking extrusion(Cheng & Hardy 2003a). Dry matter digestibility in-creased from 44% to 67% when barley was extrudedand energy digestibility increased to 70%. Barleyprotein is well digested, and carbohydrate and energyonly moderately digested by hybrid tilapia with pro-tein, carbohydrate and energy digestibility coe⁄-cients of 87%, 60% and 63% respectively (Sklan,Prag & Lupatsch 2004). Several approaches may im-prove digestibility of nutrients from barley. Linberg,Arvidsson and Wang (2003) noted small improve-ments in organic matter and starch digestibility forpiglets from barley cultivars selected for varied amy-lose to amylopectin ratios. Kaiser, Bowman, Surber,Blake and Borowski (2004) also noted cultivar di¡er-ences in digestibility of dry matter, starch and acid-detergent ¢bre utilizing rats as an experimental mod-el. Di¡erences in dry matter and energy digestibilitycoe⁄cients may be due in part to the inability ofmonogastric animals to digest glucans and the ef-fects glucans have on intestinal content viscosity(Bergh, Razdan & Aman1999). These authors deter-mined that b-glucanase supplementation of poultryfeeds can improve illeal digestibility of starch anddry matter from barley.

Phytic acid

Phytate phosphorus is another undesirable compo-nent of barley, as its ultimate fate is to be excretedand thus released in e¥uent. Barley may contain asmuch as 82% of its phosphorus as phytate phos-phorus but varieties of low-phytate barley are avail-able that will increase available P in barley (Sugiura,Raboy, Young, Dong & Hardy 1999; Overturf, Raboy,Cheng &Hardy 2003).The inclusionof microbial phy-tase also has been demonstrated to release phytatephosphorus in barley for absorption by rainbowtrout (Cheng & Hardy 2002). Phytate in barley is notdesirable, but should not limit its use in aquafeeds.



b-glucans

b-glucans occur in the bran of grasses such as barley,oats, rye and wheat, at concentrations of about 7%,5%, 2% and o1% respectively. As with protein con-tent, growing conditions and variety a¡ect b-glucancomposition (content and linkage types) of barley,and ranges from 4% to11%. The metabolic e¡ects ofglucans is not clear in ¢sh but have been demon-strated to be immunomodulatory in some species(Sealey & Gatlin III 1999). The duration of feeding b-glucan will determine its e¡ect. Prolonged feedingmay be detrimental and short-duration feeding maybe bene¢cial. Although a bodyof evidence on the me-tabolic e¡ects of b-glucan on ¢sh is building, the le-vels and feeding protocols for use of barley are stillbeing evaluated. In other livestock species, glucanshave been detrimental to growth and feed e⁄ciency.

Canola

Glucosinolates

As mentioned above, glucosinolates are compoundsfound in rapeseed/canola meal. Their hydrolysis pro-ducts are antithyroid substances. Goitrin is the mostpotent of the hydrolysis products and acts by inhibit-ing the organic binding of iodine, whereas anotherhydrolysis product, thiocyanate, lowers iodine up-take by the thyroid. Thus, glucosinolates are notthemselves harmful substances, as are their hydroly-sis products. Supplementing additional iodine to thediet does not prevent thyroid insu⁄ciency in ¢sh feddiets containing rapeseed meal. Canola meal is pro-duced from cultivars of rapeseed that have been bredto contain low levels of glucosinolates relative to le-vels found in rapeseed meal. Rapeseed meals containfrom 3% to 8% glucosinolates, whereas canola con-tains o0.2%, a level below that causing physiologi-cal e¡ects in ¢sh.

Erucic acid

Erucic acid (C22:1, n-9) is a normal constituent of ra-peseed oil, and levels in oil from standard varieties ofrapeseed range from 20% to 55% of the oil. Erucicacid is valuable for production of industrial lubri-cants from rapeseed oil, but deleterious in oils usedfor livestock and ¢sh diets. Erucic acid is cardiotoxic,causing cardiac lesions in rats, evenat low levels of1^2% of the lipid fraction. Varieties of rapeseed havebeen developed that contain reduced levels of erucicacid, and solvent extraction of rapeseed/canola

following mechanical pressing of the seeds reducesthe residual amount of erucic acid in canola meal tolevels below that required to have pathological e¡ectsin ¢sh.

Phytic acid

Asmentioned above, phytic acid is the storage formofphosphorus in all seeds, including grains and oil-seeds fromwhich products used in livestock and ¢shfeeds are produced. Levels in canola meal average4.0%, similar to that found in SBM.Total phosphorusin canola meal is 11g kg�1, and phytate phosphorusaccounts for about 8.3 g kg�1, or about 75% of the to-tal phosphorus in canola meal. Treatment with phy-tase increases total phosphorus digestibilityof canolameal to rainbow trout from 12% to 42% (Cheng &Hardy 2002).

Corn

Pigments

Corn products contain xanthophylls, a group ofyellow carotenoid pigments, and pigment levels areconcentrated in corn gluten meal compared withthe xanthophyll content of ground corn. Rainbowtrout fed diets containing corn gluten meal depositxanthophyll pigments in the muscle tissue, resultingin a yellow colour in ¢llets. For trout receiving dietslacking supplemental astaxanthin or canthaxanthin,corn gluten meal cannot be used, but in trout feddiets supplemented with these carotenoid pigmentsto produce red ¢llets, the presence of corn glutenmeal in the diet does not a¡ect ¢llet colour.

Lysine limitation

Corn is de¢cient in lysine compared with othergrains and oilseeds, and in ¢sh diets lacking proteiningredients produced from ¢sh or animal/poultry by-products, lysine is nearly always the ¢rst limitingamino acid. Thus, ¢sh diets cannot be formulated tobe amino acid su⁄cient when corn gluten meal is amajor constituent unless supplemental lysine is pro-vided. Corn gluten meal contains 1.02% lysine, com-pared with 4.7% in menhaden meal. Soybeam mealcontains 3.02% lysine and SPC contains 4.23% ly-sine. Thus, combining corn gluten meal with SBM orSPCmay result in a blend containing lysine at nearlyacceptable levels.



Cottonseed meal

Gossypol

Cottonseed meal typically contains about 400^800mg of free gossypol kg�1, a compound recog-nized for its toxicity to animals, especially in regardto reproduction. Herman (1970) indicated that inrainbow trout growth depression did not occur atgossypol concentrations lower than 290mg kg�1

diet whereas histopathological changes in the liverand kidney were noticed at 95mg kg�1. Gossypol ac-cumulated in the liver when trout were fed 1000mgof gossypol acetate kg�1 diet for several months. Leeet al. (2002) examined concentrations of gossypol iso-mers in liver, muscle, kidney, intestine and gonadsfollowing 9 months of feeding rainbow trout lowand high levels of CSM-containing diets. This studyindicated lower retention of the (� ) rather than (1)gossypol isomer. Administration of gossypol byperitoneal injections established e¡ects of watertemperature, injection dose and isomer compositionin tissues of rainbow trout that would allow estima-tion of post-prandial behaviour of gossypol and itsmajor metabolite, gossypolone (K. H. Park andK. Dabrowski, pers. comm.). Channel cat¢sh fed dietscontaining CSM at 29% of diet with an equivalentamount of gossypol acetate (0.156%) had 1.8^2.2mgof gossypol kg�1 of muscle tissue (Dorsa, Robinette,Robinson & Poe 1982). Robinson and Tiersch (1995)found that mature cat¢sh fed diets containing CSMat either 37.5% or 52% (300 or 400mg of freegossypol kg�1 diet) gained less weight than ¢sh fedeither a control diet (no CSM) or a diet with 25%CSM. Gossypol concentrations in liver were propor-tional to dietary CSM levels.Further studies should address themeans of gossy-

pol detoxi¢cation at the gastrointestinal level for pre-venting pathological changes in haematology, forinstance in tilapia Oreochromis sp. (Rinchard, Mba-hinzireki, Dabrowski, Lee, Garcia-Abiado & Ottobre2002; Garcia-Abiado, Mbahinzireki, Rinchard, Lee &Dabrowski 2005), preventing inhibitionof pepsinandtrypsin, as well as accelerating removal of gossypolvia bile and post-absorptive transport. Impaired ab-sorption of gossypol may allow complete utilizationof CSM protein, and realization of growth-promotinge¡ects, for instance, via £avonoid contributions (anti-oxidant action) or steroidogenesis, as stimulationof testosterone concentrations have been observedin rainbow trout (Dabrowski, Rinchard, Lee, Blom,Ciereszko & Ottobre 2000) and tilapia (Rinchardet al. 2002).

Quercitin

Quercetin is rapidly metabolized to isorhamnetin,which is a weaker antioxidant in animals (Manach,Texier, Morand, Crespy, Regerat, Demigne & Remesy1999) but may have oestrogenic e¡ects. Consideringthe rapid conversionof quercetin inmammalian cells(Manach, Morand, Crespy, Demigne,Texier, Regerat &Remesy 1998), the prolonged presence in its parentform in ¢sh diets may be an advantageous assetvis-a-vis its pharmacological e¡ects.Both quercetin and isorhamnetin exist almost ex-

clusively as glucuronides/sulphates (Piskula & Terao1998; Erlund, Kosonen, Alfthan, Maenpaa, Perttu-nen, Kenraali, Parantainen & Aro 2000;Wittig, Her-derich, Graefe & Veit 2001; Hou, Chao, Ho,Wen&Hsiu2003). In preliminary studies, low levels of quercetinand isorhamnetinwere occasionally detected in liverand body tissues of tilapia reared for15weeks but notsupplemented with quercetin. The source of querce-tin in this case might have been algae growing ontank walls, but this was not ascertained, thus imply-ing that Nile tilapia has aunique ability to absorb andaccumulate quercetin. It is not yet known whetherquercetin absorption is much more e⁄cient in tilapiacompared with other ¢sh, considering the relativelyhigh levels of quercetin in CSM (when supplementedat 54%) and the failure to detect quercetin in rain-bow trout blood plasma (K. J. Lee & K. Dabrowski,pers. comm.).

Peas/lupins

Alkaloids

The predominant antinutrients for ¢sh in lupins arethe quinolizidine alkaloids (Duke 1981). There areover 300 species of the genus Lupinus (L.), of whichmany contain high levels of alkaloids (‘bitter’ vari-eties). These alkaloids, a class of bitter-tasting com-pounds, make the seed unpalatable and sometimestoxic to animals.There are several di¡erent alkaloids present in lu-

pins, and for cattle, Duke (1981) reported ‘lupin poi-soning’ due to quinolizidine alkaloids or their N-oxides, and ranked the toxins in order of increasingimportance: D-lupanine, sparteine, lupanine, spathu-latine and hydroxylupanine. As the alkaloids are notheat labile, they have traditionally been removed bysoaking the seed in water; however, plant breedershave selected strains of alkaloid-free or ‘sweet’ lupins,which could be directly consumed by livestock (Duke



1981). Commercially available strains of lupins thatare commonly used for livestock are ‘sweet’ typessuch as white lupin (L. albus Linnaeus), yellow lupin(L. luteus Linnaeus) and blue or narrow-leafed lupin(L. angustifolius Eastwood) (Waldroup & Smith 1989;Nelson & Delane1990).These strains are usually verylow in alkaloid (20mg kg�1) when compared with‘bitter’ seeds that have a much greater alkaloid level(15000^22000mg kg�1) (Luckett 2004). Low alka-loid strains of lupins do not appear to present palat-ability problems to ¢sh. Glencross, Carter et al.(2004) fed diets containing up to a 50% inclusionlevel of yellow lupin kernel meal for 42 days to rain-bow trout and found no evidence of reduced feedintake. However, growth performance was limitedwhen lupin meal was included in the diets beyond30% (Glencross, Carter et al. 2004).

Oligosaccharides

Moderate levels of oligosaccharides are present inlupins and ¢eld peas (Duke 1981; Hickling 2003).The oligosaccharide component of lupins and ¢eldpeas are mainly composed of a-galactosides (Guillon& Champ 2002), which are indigestible by most spe-cies of ¢sh (Stickney & Shumway 1974; Kuz’mina1996). The oligosaccharide levels of lupins and ¢eldpeas may vary between strains but they are approxi-mately 8% (predominantly stachyose 3%, sucrose2%, verbascose 2% and ra⁄nose 0.8%) and �5.0%(predominantly sucrose 2% and verbascose 1.5%and ra⁄nose 0.5%), respectively, of the total seed ona dry basis (Guillon & Champ 2002; Hickling 2003).The oligosaccharide levels of lupins and ¢eld peas

are low compared with levels of 15% found in SBM(Russett 2002), and may present less of a problem.The oligosaccharide component of SBM have beenlinked with reduced digestibility, growth perfor-mance (Refstie et al. 1998) and the occurrence ofSBM-induced enteritis in several salmonid ¢sh spe-cies (vanden Ingh et al.1996; Bureau et al.1998).Whilethere was evidence of reduced digestibility due to theoligosaccharide component of lupin (L. angustifoilius)meal for rainbow trout (Glencross, Boujardand et al.2003), there was no evidence of a condition similarto SBM-induced enteritis in rainbow trout fed dietscontaining 50% inclusion level of yellow lupin kernelmeal for 42 days (Glencross, Evans et al. 2004).

Lysine limitation

For lupins and ¢eld peas, both lysine and methionineare limiting amino acids (Table 3). Therefore, diets

based on lupins or pea seeds will require added sup-plemental lysine and methionine. Ultimately, for theinclusion of lupins and ¢eld peas into ¢sh diets itwould be desirable to develop strains higher in lysineand methionine, remove the oligosaccharide compo-nent of both legumes and ensure selection of lowalkaloid strains of lupins.

Wheat

Lysine limitation

Lysine is the most limiting indispensable amino acidin wheat for various aquatic species.Wheat breedinghas long emphasized improvements in grain and£our quality with steady improvements in the glutenstrength of cultivars grown by farmers (Cox, Shog-ren, Sears, Martin & Bolte1989; Souza,Tyler, Kephart& Kruk 1993). Because of the importance of doughrheology in the end-use of bread wheats, selectionhas emphasized functional di¡erences in seed sto-rage proteins rather than selection to enhance nutri-tional composition, such as elevating the percentageof lysine and methionine. Wheat, like other relatedcereals, has only limited variation for seed aminoacid composition. Transgenic modi¢cation of seedstorage proteins is possible but would be likely to en-counter sti¡ opposition from the grain industry thathas chosen to avoid transgenic technology for thenear future.

Researchable issues and approaches toincrease use of plant products inaquafeeds

Processing and techniques to enhancenutrient digestibility/nutritional value andquality

Feed-processing methods have been used for manyyears to improve the physical characteristics and nu-tritional value of feeds, and the e¡ects of di¡erentmanufacturing methods are well known (Booth,Allan & Warner-Smith 2000; Cheng & Hardy 2003b;Venou, Alexis, Fountoulaki, Nengas, Apostolopoulou& Castritsi-Cathariou 2003). Cooking extrusion, forexample, increases carbohydrate digestibility (Allan& Booth 2004) and produces a more durable pelletthat can be controlled to make the pellet £oat or sink(Barrows & Hardy 2001). Pre-processing refers totreatment of speci¢c ingredients before mixing theingredients together into a complete feed. Routine



pre-processing of SBM reduces the level of trypsininhibitors. Other antinutrients (i.e. oligosaccharides,¢bre, phytate), however, are still present in SBM, andother plant ingredients, after traditional pre-proces-sing has occurred and thus can have negative e¡ectson ¢sh (Olsen, Myklebust, Kryvi, Mayhew & Ringo2001). To increase the nutritional value of plant-derived ingredients for ¢sh, pre-processing of ingre-dients using both biological enhancement andmechanical modi¢cation are being explored.Biological enhancement involves the use of micro-

organisms to modify the chemical composition of theingredient.Yeast, bacterial and fungal fermentationshave been investigated and current results demon-strate great potential for removing antinutrients andadding essential nutrients such as protein and aminoacids (Mukhopadhyay & Ray1999; Bairagi, Gosh, Sen& Ray 2004). The overall goal of these processes is toincrease protein concentration and decrease thelevels of antinutrients. Removal of the exotic or fer-mentable carbohydrates and phytate is well withinthe capabilities of this approach (Ng, Lim, Lim & Ibra-him 2002; Skrede, Storebakken, Skrede, Sahlstrom,Sorensen, Shearer & Slinde 2002). Microorganismsconsume and convert carbohydrates from unavail-able energy into cell mass that serves as a digestiblesource of both protein and energy.The type of micro-organisms and plant substrates used in the fermenta-tion will determine the nutrient and antinutrientpro¢le of the ¢nal product. The cost of producing bio-logically enhanced plant products will be the pri-mary factor a¡ecting their use in future aquafeedformulations.Mechanical modi¢cation of plant-derived ingredi-

ents is an extensionof current pre-processing techni-ques. These methods can be as simple as de-hullingbarley before grinding and mixing to decrease the ¢-bre content and thus increase protein content (Hardy& Barrows 2000). The more complex pre-processingmethod of air classi¢cationuses air pressure and par-ticle density to separate dense protein particles fromthe lighter carbohydrate fractions and results in aproduct with higher nutrient levels (Thiessen, Camp-bell & Tyler 2003;Thiessen, Campbell & Adelizi 2003).Rice protein is currently being commercially concen-trated from approximately 12^70% protein using airclassi¢cation. The protein concentration of barleycan be doubled from12% to 25% using air classi¢ca-tion and this product is also currently commerciallyavailable. Other methods to increase nutrient and de-crease antinutrients are also being developed andevaluated (Kim, Flores, Chung & Bechtel 2003). Pro-

tein concentrates and isolates from wheat, soy andcanola are commercially available, but are currentlyrelatively expensive for use in aquafeeds. As with bio-logically enhanced proteins, the ¢nal cost of the pro-duct will determine the e¡ectiveness and commercialapplication of the process.

Augmenting nutritional quality throughgenetic manipulations of plants

The nutrient and antinutrient pro¢les of plant pro-ducts are currently not ideal for aquafeeds. Fortu-nately, it is possible to change speci¢c traits of grains,such as protein and oil content, through breedingand genetic manipulation. Genomic and genetic stu-dies have determined the molecular mechanisms in-volved with the expression of di¡erent nutrient traits.Knowledge from those studies, coupled with moderntechnologies of genetic and molecular biology, pro-vides powerful tools for breeders and geneticists tomanipulate plants more quickly and e⁄ciently toachieve desired traits. In the last decade, genetic ma-nipulations have changed, as have the potential tofurther improve several speci¢c nutrients in seeds.Examples of changes in several important antinutri-ents and nutrients are listed below.

Low phytic acid mutations

Phytic acid is the major form of phosphorus, repre-senting 60^80% of the total phosphorus in matureseeds and is not usable by monogastric animals.Low phytic acid (up to 75% reduction in seed) buthigh free-phosphorus genetic materials have beenidenti¢ed in barley, maize, wheat, rice and soybeanthrough mutations (Dorsch, Cook,Young, Anderson,Bauman,Volkmann, Murthy & Raboy 2003).This lowphytic acid germplasm is being used in breeding pro-grammes to produce commercially viable cultivarsthat will help reduce phosphorus discharge fromani-mal production units. The ¢rst low phytic acid muta-tions in wheat did reduce the amount of phytic acidby 30% or more while increasing the concentrationof inorganic phosphorus in the grain (Guttieri et al.2004). However, wheat appears to di¡er from othercereals in having signi¢cant genetic variation amongadapted cultivars for the enzyme phytase, which de-grades phytic acid. Previous studies have reported agreater than threefold di¡erence in phytase activityamong limited sub-sets of high-yielding cultivars(Okot-Kotber,Yong, Bagorogoza & Liavoga 2003). Un-published work at the University of Idaho has found



similar activities among spring wheat cultivars andbreeding lines adapted to the Northern United States(M. J. Guttieri & D. Peterson, pers. comm.). Other ge-netic manipulations in wheat mineral compositioninclude selection for high iron and high calciumconcentrations.

High lysine genetic materials

Lysine is an EAA for animals that is often limiting inplant feedstu¡s. Crops with high lysine content are ofgreater value, but plants in the grass family generallyhave low lysine content. However, genetic studieshave identi¢ed high lysine mutations in corn andbarley where the lysine level is increased up to150%. More recent studies have found the modi¢ergenes of the well-known opaque-2 phenotype inmaize. Plants carrying the modi¢er genes producenormal corn seeds with the same high level of lysineas that in the opaque-2 mutation (Gibbon & Larkins2005). Equally exciting progress has occurred in thesuccessful manipulation of the genetic pathwayof ly-sine catabolism for high lysine accumulation in Ara-bidopsis by disruption of the gene in that pathway(Stepansky,Yao,Tang & Galili 2004).The same geneticapproach could be adapted in all other crops for highlysine production.

b-glucan manipulation

Researchers at the USDA/ARS National Small GrainsGermplasm Research Facility (Aberdeen, ID, USA)have obtained barleygermplasmwith b-glucan levelsranging from 1.5% to over 10.0% through breedingselection and chemical mutation. These materialsare being incorporated into elite lines for variouslevels of b-glucan. Both low and high b-glucanlines may satisfy the needs of direct use and industryprocessing by-products for aquafeed production.

High-oil cereal crops

Quantitative trait loci controlling oil content havebeen identi¢ed in cereal maize in which oil contentcould approach 20% by selection (Laurie, Chasalow,LeDeaux, McCarroll, Bush, Hauge, Lai, Clark, Roche-ford & Dudley 2004). Quantitative trait loci identi¢edin maize may help oil improvement in other cerealcrops as well.

Starch-structure manipulation

The level of di¡erent starch structures such as amy-lose and amylopectin can be changed according to

the requirement of aquafeeds through better under-standing of the genes in£uencing starch biosyntheticpathways (Morell & Myers 2005). In wheat grain,modi¢cation of starch to lower amylose content toproduce amylose-free wheat starch (sometimescalled ‘waxy’) is genetically possible through accu-mulationof threemutant forms of the granule-boundstarch synthase gene. The ¢rst germplasm lines withamylase-free starch are widely available (Graybosch,Souza & Berzonsky 2003) and the ¢rst commercialamylose-free starch cultivars are now reaching themarketplace. Development of high-amylose starchwheat cultivars is also possible (Yamamori, Fujita,Hayakawa & Matsuki 2000). However, cultivardevelopment in this area is less advanced than thedevelopment of cultivars with amylose-free starch.Modi¢cations of starch can alter the functional prop-erties and the digestibility of the starch.

Increased micronutrient content

Genes regulating micronutrients such as antioxida-tive compounds (like vitamin E) have been identi¢edin Arabidopsis (Capell & Christou 2004). That discov-erymayallowenhancement of such traits in crops forthe possible bene¢t of ¢sh.

Enhancing utilization by genetic selection of¢sh

Evaluation of genotype by diet interactions in aqua-culture species for speci¢c dietary components hasonly recently begun on a limited basis. Initial studieshave examined such species as sea bream, rainbowtrout and Atlantic salmon (Vivas, Sanchez-Vazquez,Garcia & Madrid 2003; Overturf, Bullock, LaPatra &Hardy 2004). Because of the high-protein diet thesespecies consume in the wild, commercial diets haverelied heavily on ¢sh meal and ¢sh oil as protein andenergy sources. Other omnivorous ¢sh species suchas tilapia and cat¢sh have demonstrated a greaterproclivity for utilizing plant feedstu¡s and carbo-hydrate for energy but little research has beenperformed on these species in regard to alternativedietary selection. Furthermore, a large number ofaquaculture species have only recently become do-mesticated and as such most of the initial researchwith these species has involved general determina-tion of their dietary requirements. Because of thegreat diversity of species that may be cultured, thephysiologyandmetabolismof these organisms variessubstantially. As such, this necessitates that every



species be evaluated individually for its optimal nu-trient requirements.When these are known, evalua-tion of diversity within the species for utilization ofspeci¢c nutrient components can be performed andestimates made regarding improvement by a selec-tion programme. However, in other animal systems,the e¡ects of dietary changes on speci¢c physiologi-cal parameters have been studied. For example, dietby strain di¡erences has been shown in cattle andpoultry (Wahstrom,Tauson & Elwinger 1998; Roche,Kolver, deVeth &Napper 2001; Co¡ey, Simm, Oldham,Hill & Brotherstone 2004; Jones, Zaczek, MacLeod &Hocking 2004) and this has been more speci¢callylinked to individual genes in work performed withmurine and ¢sh models such as medaka, Oryziaslatipes (Corva & Medrano 2000; Kaput, Klein,Reyes, Kibbe, Cooney, Jovanovic,Visek & Wol¡ 2004;Hostetler, Collodi, Devlin & Muir 2005).In aquaculture, there have been very few studies

attempting to discern the e¡ects of genotype on nu-trient utilization. Initial studies have focused on re-placement of ¢sh meal protein with that from plants,typically using re¢ned feedstu¡s. Hence, the proteincomponent as well as antinutritional factors andother factors have been altered, sowhat is being eval-uated at this level is the lack of hormones and otherby-products found in ¢sh meal with that of the repla-cement protein. This severely limits our ability tointerpret the e¡ect of unre¢ned materials on the ani-mal’s metabolism and physiology. All other dietarycomponents relating to amino acid levels, nitrogenlevels, etc., are held constant. Also, the high cost ofproduction of re¢ned diets makes them of limited va-lue for aquaculture feeding practices. Of greater in-terest is the determination of whether carnivorous¢sh that have been evolutionary selected to utilizeprotein as their main energy source can be selectedfor improved utilization of plant material. In the past,diet studies were roughly correlated by relatinggrowth indices such as weight gain, feed e⁄ciency,growth rate, digestibility and proximate analysis ofwhole ¢llets to variable changes of dietary compo-nents.With the recent impetus of genomics, and theincreasing amount of sequence information avail-able, an improved understanding of cellular signal-ling, and the role of genes and proteins in pathways,now several research groups are beginning to char-acterize what is occurring in the animal not onlyphysiologically but also at the cellular and genomiclevels.Several vertebrate genomes have been completely

sequenced and this allows for the retrieval of homo-

logous sequences from previously uncharacterizedgenomes. With the availability of this information,the correlations of biochemical information withgenetic and molecular data will be very useful inproviding improved insight into the functions of un-known genes or systems responses to nutrientchanges. This research is being pursued even moreintensively via the use of microarrays where scien-tists can now screen for the expression of thousandsof genes in single experiments.This will revolutionizehow nutritionists evaluate diets in that now they canrelate precise dietary changes with speci¢c physiolo-gical pathways that are being modulated by nutri-tional di¡erences.Future work requiring genomic information will

focus on the current most highly characterized spe-cies such as Atlantic salmon, channel cat¢sh andrainbow trout and may incorporate informationgained from other highly developed research speciessuch as zebra¢sh, Brachydanio rerio, fugu ¢sh,Takifu-gu £avidus and medaka.

Optimizing bioactive compounds

Several plant feedstu¡s contain bioactive compoundsthat may have positive or negative e¡ects on aquaticanimals and thus their concentrations in aquafeedsmust be adjusted accordingly. For example, cotton-seed is a source of many phytochemicals such as £a-vonoids. Quercetin, a common £avonoid in plants,was measured at concentrations from 1320 to1560mg kg�1 in three di¡erent cottonseed prepara-tions (Lee et al. 2002). Many positive functions havebeen attributed to quercetin upon its intake by hu-mans and it has also drawn attention for decadesdue to its bene¢cial e¡ects as an antioxidant agent in¢sh oil storage (Martinez Nieto, Garrido Hoyos, Ca-macho Rubio, Garcia Pareja & Ramos Cormenzana1993). Owing to such antioxidant activities, this£avonoid also has been proven e⁄cacious againstvarious pathological conditions such as cancer, car-diovascular disease and other oxidative stresses (For-mica & Regelson 1995). Despite the ubiquitousoccurrence of quercetin in feedstu¡s such as CSM(Shahidi & Naczk 1995), and thus possible ingestionby ¢sh (Rinchard, Lee, Czesny, Ciereszko & Dabrows-ki 2003), little is known of its nutritional role in for-mulated aquafeeds. The only available reportindicates that quercetin is not overly toxic to rainbowtrout when fed for several months at high levels(Plakas, Lee & Wolke 1985). Some other reports,



however, claimed high toxicity of quercetin to thereproductive system of ¢sh upon external exposure(Weber, Kiparissis, Hwang, Niimi, Janz & Metcalfe2002).Other bioactive compounds such as many of the

proteins found in SBM must be characterized as tothe immunological basis of their food-intolerance ac-tivity to determinewhether the same proteins induceintolerance in diverse species of ¢sh. If soy proteinsare identi¢ed that induce immunological responsein ¢sh species of commercial interest, varieties lack-ing those proteins could be stacked with other traitsdesirable to the aquaculture industry. Speci¢c stepsto accomplishing these goals will be: (1) identify spe-ci¢c proteins that induce soybean-food intolerance in¢sh; (2) determine whether the adverse reactions aredue to the physical presence of the proteins or depen-dent on their biological activity; and (3) determinethe extent of intolerance-response commonalityamong farmed ¢sh species. On the basis of thisapproach, a strategy to mitigate food intolerance byaltering seed protein content maycreate the opportu-nity to use renewable plant-based feedstu¡s as substi-tutes for ¢sh meal. To accomplish this, teamresources including laboratories with experience in¢sh growth and physiologywill be needed alongwithother laboratories experienced in analyzing adverseimmunological reactions and having the capabilityto modify the protein component of the feedstu¡.

Using pre- and probiotic compounds toenhance utilization

It has been documented in a number of food animalsthat their gastrointestinal microbiota plays impor-tant roles in a¡ecting the nutrition and health of thehost organism. Thus, various means of altering theintestinal microbiota to achieve favourable e¡ectssuch as enhancing growth, digestion, immunity anddisease resistance of the host organism have beeninvestigated in various terrestrial livestock as well asin humans. Dietary supplementation of prebiotics,which are classi¢ed as non-digestible food ingredi-ents that bene¢cially a¡ect the host by stimulatinggrowth and/or activity of a limited number ofhealth-promoting bacteria such as Lactobacillus andBi¢dobacter spp. in the intestine, while limitingpotentially pathogenic bacteria such as Salmonella,Listeria and Escherichia coli, have been reported tofavourablya¡ect various terrestrial species. However,such information is extremely limited to date for

aquatic organisms. E¡ects of probiotics, de¢ned aslive microbial-feed supplements, on gastrointestinalmicrobiota have been studied in some ¢sh, but theprimary application of microbial manipulations inaquaculture has been to alter the composition of theaquatic medium. In general, the gastrointestinal mi-crobiota of ¢sh including those produced in aquacul-ture has been poorly characterized, especially theanaerobic microbiota. Therefore, more detailedstudies of the microbial community of cultured ¢share needed to potentially enhance the e¡ectivenessof prebiotic and probiotic supplementation.

Monitoring e¡ects of plant feedstu¡s on ¢shproduct quality and consumer health

Success or failure in anycommercial aquaculture en-deavour ultimately depends on repeat purchasing byconsumers, whether at the retail, wholesale or foodservice level. Aquatic foods market research consis-tently indicates that product quality is the singlemost important attribute e¡ecting ¢sh-purchasingbehaviour. Given the recent onslaught of negativemedia attention to farm-raised ¢sh, consumers arealso more focused on the potential health e¡ects as-sociated with ¢sh consumption. As we move forwardtoward increasing the use of plant products in aqua-feeds, the e¡ects of these newly formulated diets on¢sh product quality and consumer health must beevaluated.Although quality is a complex issue, for the pur-

poses of this paper it will refer primarily to the colour,£avour, texture and nutrient composition of the ¢shproduct. Numerous studies have reported that all ofthe previous attributes, as well as shelf-life stabilityof ¢sh ¢llets, can be a¡ected by diet composition. Anoverwhelming majority of the studies, particularlythose that have included the use of sensory methodsto evaluate quality, have focused on salmonids,primarily rainbow trout and Atlantic salmon. Non-salmonid species sporadically evaluated for dietaryin£uence on sensory quality include Gulf sturgeon,Acipenser sp. Linnaeus, hybrid striped bass, whiteamur, Ctenopharyngodon idella Valenciennes, turbot,barramundi, perch, tilapia and channel cat¢sh. Fla-vour, colour, odour and/or texture were signi¢cantlya¡ected by diet in about half the studies reported.Given the physiological, nutritional, environmentaland compositional di¡erences among aquacultured¢n¢sh, conclusions reached about the product qual-ity of one species cannot be automatically applied toanother. Further research is needed to clarify the



in£uence of dietary ingredients on quality of eachaquaculture species of interest.Three di¡erent methods of sensory analysis have

typically been used to characterize the organolepticquality of ¢sh ¢llets in the studies indicated above:descriptive analysis, di¡erence testing and consu-mer-acceptability testing. The most sophisticated ofthese, descriptive analysis, uses trained sensory pa-nelists to pro¢le individual £avour, odour and texturedescriptors in the ¢sh, and to rate them for theirintensity. Di¡erent £avour attributes, such as nutty,sweet, ¢shy, painty or grassy, have been used to de-scribe the e¡ects of dietary ingredients on the speci¢c£avour characteristics of ¢sh from di¡erent dietarytreatments. Although this type of information canbe extremely useful in the later stages of product-quality evaluationand modi¢cation, for the purposesof initial evaluation of plant-based ingredients inaquafeeds, untrained consumer panels should be uti-lized.Trained panelists have been reported to discerndi¡erences among samples that average consumerscannot. Research in this ¢eld should attempt to an-swer two questions that have important marketingrami¢cations: (1) can the average consumer tell thedi¡erence between ¢llets from ¢sh reared on theplant-based diet versus a ¢sh meal control diet, and(2) does the average consumer prefer ¢sh from onedietary treatment over the other?While numerous studies have evaluated the e¡ects

of various alternative plant proteins on ¢sh growthand feed e⁄ciency, relatively few have monitored thedietary in£uence on ¢sh quality. However, a reviewof15 studies, which investigated the e¡ects of variousplant proteins, including SBM, SPC, soy £our, corngluten, wheat gluten, peanut meal, canola meal, ex-truded peas, lupin seed meal and CSM, on ¢llet qual-ity, indicates that product quality was signi¢cantlyin£uenced by protein source in about 40% of the stu-dies. Colour, as measured by instrumental and sen-sory methods, was the attribute most in£uenced byprotein source.Texture and £avourwere signi¢cantlya¡ected by protein source in some cases, particularlywhen large substitutions were made. However, otherstudies inwhich 30^60% of the diet was composed ofplant protein meals reported no di¡erences in consu-mer acceptability among ¢llets. Clearly, the e¡ects ofplant protein meals, particularly those subjected tovarious degrees of concentration and/or isolation,on product quality require further investigation.Two important areas related to consumer health

that should be addressed with regard to utilizationof plant products in aquafeeds are o-3 fatty acid level

and possible PCB contamination in ¢llets. Becausefatty acid composition of ¢sh ¢llets is related to thefatty acid composition of the diet, ¢sh fed primarilyplant-based diets contain lower levels of o-3 fattyacids. Given that consumers are becoming more in-terested in the health bene¢ts of these fatty acids, fu-ture research should include evaluating the e¡ects of¢nishing diets rich ino-3 fattyacids on fattyacid pro-¢le and sensory quality of the ¢llets. As PCB contam-ination of farmed ¢sh arises primarily from ¢sh oil,methods need to be developed that enrich the o-3content of the ¢llet without increasing contaminantlevels. Although the risks of organochlorine contam-ination from farmed ¢sh are open to debate, consu-mers are aware of this issue, which may a¡ect theirpurchasing behaviour.

Enhancing palatability of plant feedstu¡s

Fish show distinct taste preferences in basic studiesof taste and smell. In a recent review paper (Kasum-yan & Doving 2003), the state of the art is sum-marised as follows: ‘Fish taste preferences are highlyspecies speci¢c, and the di¡erences among ¢sh spe-cies are apparent when comparing the width andcomposition of spectra for both the stimulants andthe deterrents.What is evident is that there is a strongsimilarity in the taste preferences between geogra-phically isolated ¢sh populations of the same species,and that taste preferences are similar in males andfemales, although at the individual level it may varydramatically among conspeci¢cs. What is note-worthy is that taste responses are more stable andinvariable for highly palatable substances than forsubstances with a low level of palatability. There is agood correspondence between development of thegustatory system in ¢sh ontogeny and its ability todiscriminate the taste properties of food items. Thereis also a correspondence between oral and extraoraltaste preferences for a given species; however, there isno correlation between smell and taste preferences.Taste preferences in ¢sh show low plasticity (in rela-tion to the diet), appear to be determined geneticallyand seem to be patroclinous. Fish-feeding motivationand various environmental factors like water tem-perature and pollutants such as heavy metals andlow pH water may shift ¢sh taste preferences. Com-parisons between bioassay and electrophysiologicaldata show that palatability is not synonymous withexcitability in the gustatory system’.The methodologies and expertise gained in basic

studies should be used to characterize the taste



elements of plant feedstu¡s and how processing al-ters palatability. Research regarding palatability ofvarious feedstu¡s may indicate why feed intake isoften reduced when certain feedstu¡s are includedin ¢sh diets and suggest how processing may be con-ducted to optimize palatability. The fact that palat-ability is the sum of many dietary characteristicsand that there is a strong interaction between tasteand nutritional quality should be re£ected in theexperimental designs. Such interactions have notgained much attention in research so far, either in¢sh nutrition or in the nutrition of other animals.

Metabolomics

Metabolomics is the studyof themetabolic pro¢le of agiven cell, tissue, £uid, organ or organism at a givenpoint in time. The metabolome represents the endproducts of gene expression. Thus, mRNA gene ex-pression data and proteomic analyses do not tell thewhole story of what might be happening in a cell(mRNA gene transcripts do not necessarily re£ecteither concentration or activity of a gene product ina cell) and do not account for post-translational mod-i¢cations. Of all the feedstu¡s described in this docu-ment, soybean is the most widely studied with itschemical composition of the major nutrient compo-nents as well as the NSPhaving been described in de-tail and variability being well characterized. Also thecontents of a number of biologically active compo-nents, some of which belong to the rather ill-de¢nedgroup of antinutrients, have been described to a cer-tain extent. However, awide range of components arestill unknown. The possibility that some of thesecomponents may have biological e¡ects of positive ornegative character is very likely. Despite investigationof the adverse e¡ects of soybeans in salmonids fortwo decades, the causing agent(s) has not beenfound. The component may well be among the uni-denti¢ed compounds. Developments in instrumenta-tion and chemical approaches over the last few yearshave greatly improved.It appears that to describe and predict the nutri-

tional value of soybeans and other plant feedstu¡son the one hand, and possible adverse antinutrientand toxic e¡ects on the other hand, it is imperativeto gain detailed chemical informationwith regard tothe secondary metabolites (Becker & Makkar 1999;Francis et al. 2001; Overturf et al. 2003; Pezzato,deOliveira, Dias, Barros & Pezzato 2004; Ulloa, vanWeerd, Huisman & Verret 2004). Metabolic pro¢lingand the identi¢cation of single volatile, semi-volatile

and £avour compounds may lead the way to un-known compounds that might impact digestibility ofthese products and in£uence the physiology, growthand health of farmed ¢sh (Johnsen & Dupree 1991;Grigorakis,Taylor & Alexis 2001). The analytes of in-terest likely belong to the following chemical groups:aliphatic hydrocarbons and their derivatives, ter-penes, phenols, N- and S-containing compounds(e.g. alkaloids, glucosinolates) (Rohlo¡ 2003, 2004).Analytical approaches have been reviewed byWilkes,Conte, Yonkyoung, Holcomb, Sutherland and Miller(2000) and Bino, Hall, Fiehn, Kopka, Saito, Draper,Nikolau, Mendes, Roessner-Tunali, Beale,Trethewey,Lange,Wurtele and Sumner (2004).A combination of analytical platforms (e.g., GC-

MS, LC-MS-MS) may be used to evaluate the meta-bolic composition of various plant feedstu¡s. Thisapproach can be used interactively to evaluate thecomposition of these feedstu¡s and fractions withdi¡erential levels of antinutrients and to identify spe-ci¢c antinutrients. These plant feedstu¡s could befractionated by extractions and preparatory HPLC,and the resultant fractions could be tested forantinutrient activity (in collaboration with ¢sherybiologists), metabolic composition and protein com-ponents. Statistical analysis would be used to identifymetabolic correlates of antifeedant activity. Antifee-dant fractions could be further fractionated and theresultant fractions tested for antinutrient activityand metabolic composition.

Conclusions and recommendations

On the basis of the information reviewed in thisdocument, speci¢c strategies and a variety of techni-ques have been identi¢ed to optimize the nutritionalcomposition of plant feedstu¡s and limit potentiallyadverse e¡ects of bioactive compounds. A coordi-nated research e¡ort needs to be developed throughstrategic planning to further evaluate and re¢nevarious means of improving the nutritional valueof plant feedstu¡s for increasing their use in aqua-culture. Such e¡orts will result in reducing thedependence on animal feedstu¡s and increasing thesustainability of aquaculture.

References

Alford B.B., Liepa G.U. & Vanbeber A.D. (1996) Cottonseedprotein: what does the future hold? Plant Foods HumanNutrition 49,1^11.



Allan G.L. & Booth M.A. (2004) E¡ects of cooking extrusionprocessing on digestibility of peas, lupins, canola mealand soybean meal in silver perch Bidyanus bidyanus(Mitchell) diets. Aquaculture Research 35,981^991.

Allan G.L. & Rowland S.J. (1994) The use of Australianoilseeds and grain legumes in aquaculture diets. In:The Third Asian Fisheries Forum (ed. by L.M. Chou, A.D.Munro, T.J. Lam, T.W. Chen, L.K.K. Cheong, J.K. Ding,K.K. Hooi, H.W. Khoo,V.P.E. Phang, K.F. Shim & C.H. Tan),pp.667^670. Asian Fisheries Society, Manila, Philippines.

Allan G.L., Parkinson S., Booth M.A., Stone D.A.J., RowlandS.J., Frances J. & Warner-Smith R. (2000) Replacement of¢shmeal in diets for Australian silver perch, Bidyanus bi-dyanus: digestibility of alternative ingredients. Aquacul-ture186, 293^310.

Almirall M., Francesch M., Perez-Vendrell A.M., Brufau J. &Esteve-Garcia E. (1995) The di¡erences in intestinal visc-osity produced by barley and �-glucanase alter digestaenzyme activities and ileal nutrient digestibilities morein broiler chicks than in cocks. Journal of Nutrition 125,947^955.

Anonymous (1992) Raw Material Compendium: A Compila-tion of Worldwide Data Sources, 1st edn. Novus Interna-tional, St. Louis, MO, USA,511pp.

Bairagi A., Gosh K.S., Sen S.K. & RayA.K. (2004) Evaluationof the nutritive value of Leucaena leucocephala leaf meal,inoculated with ¢sh intestinal bacteria Bacillus subtillisand Bacillus circulans in formulated diets for roho, Labeorohita (Hamilton) ¢ngerlings. Aquaculture Research 35,436^446.

Bakke McKellep A.M., Press C.M., Baeverfjord G., KrogdahlK. & LandsverkT. (2000) Changes in immune and enzymehistochemical phenotypes of cells of the intestinal muco-sa of Atlantic salmon (Salmo salar L) with soybean-mealinduced enteritis. Journal of Fish Disease 23,115^127.

van Barneveld R.J. (1999) Understanding the nutritionalchemistry of lupin (Lupinus spp.) seed to improve livestockproduction e⁄ciency. Nutrition Research Reviews 12,203^230.

Barrows F.T. & Hardy R.W. (2001) Nutrition and feeding. In:Fish Hatchery Management (ed. by G. Wedemeyer), 2ndedn, pp. 483^558. JohnWiley & Sons, NewYork, NY, USA.

Becker K. & Makkar H.P.S. (1999) E¡ects of dietary tannicacid and quebracho tanninnext term on growth perfor-mance and metabolic rates of common carp (Cyprinuscarpio L.). Aquaculture175,327^335.

Bedford M.R. & Classen H.L. (1992) Reduction of intestinalviscosity through manipulation of dietary rye and pento-sanase concentrations is e¡ected through changes in car-bohydrate composition of the intestinal aqueous phaseand results in improved growth rate and food conversione⁄ciency of broiler chicks. Journal of Nutrition 122,560^569.

Bennetau-Pelissero C., Breton B., Bennetau B., Corraze G.,LeMenn F., Davail-Cuisset B., Helou C. & Kaushik S.J.(2001) E¡ect of genistein-enriched diets on the endocrine

process of gametogenesis and on reproduction e⁄ciencyof the rainbow trout Oncorhynchus mykiss. General andComparative Endocrinology121,173^187.

Bennetau-Pelissero C., Latonnell K.G., Lamonthe V., Shin-karuk-Poix S. & Kaushik S.J. (2004) Screening for estro-genic activity of plant and food extracts using in vitrotrout hepatocyte cultures. Phytochemical Analysis 15,40^45.

Bergh M.O., Razdan A. & Aman P. (1999) Nutritional in£u-ence of broiler chicken diets based on covered normal,waxy and high amylose barleys with or without enzymesupplementation. Animal Feed Science and Technology 78,215^226.

B�verfjord G. & Krogdahl —. (1996) Development and re-gressionof soybeanmeal induced enteritis in Atlantic sal-mon distal intestine. A comparison with the intestines offasted ¢sh. Journal of Fish Disease19,375^387.

Bino R.J., Hall R.D., Fiehn O., Kopka J., Saito K., Draper J., Ni-kolau B.J., Mendes P., Roessner-Tunali U., Beale M.H.,Trethewey R.N., Lange B.M.,Wurtele E.S. & Sumner L.W.(2004) Potential of metabolomics as a functionalgenomics tool.Trends in Plant Science 9, 418^425.

Booth M.A., Allan G.L. & Warner-Smith R. (2000) E¡ects ofgrinding, steam conditioning and extrusion of a practicaldiet on digestibility andweight gain of silver perch, Bidya-nus bidyanus. Aquaculture182, 287^299.

Bregitzer P. (2005) ARS barley research at Aberdeen; progressin research, product development and goals. Proceedings ofthe 35th Barley Improvement Conference, Charleston,SC, pp. 48^53.

Bureau D.P., Harris A.M. & Cho C.Y. (1998) The e¡ects of pur-i¢ed alcohol extracts from soy products on the feed intakeand growth of Chinook salmon (Oncorhychus tsha-wytscha) and rainbow trout (Oncorhynchus mykiss).Aquaculture161, 27^43.

Bushuk W. & Wrigley C.W. (1974) Proteins: composition,structure and function. In:Wheat: Production and Utiliza-tion (ed. by G.E. Inglett), AVI,Westport, CT, USA,119pp.

Cai Y. & Burtle G.J. (1996) Methionine requirement of chan-nel cat¢sh fed soybean meal-corn-based diets. Journal ofAnimal Science 74,514^521.

Capell T. & Christou P. (2004) Progress in plant metabolicengineering. Current Opinion in Biochemistry 15, 148^154.

ChenJ.C. & Chen K.-W. (1997) Oxygenuptake and ammonia-N excretion of juvenile Penaeus japonicus during depura-tion following one-day exposure to di¡erent concentra-tions of saponin at di¡erent salinity levels. Aquaculture156,77^83.

Cheng Z.J. & Hardy R.W. (2002) E¡ect of microbial phytaseon apparent nutrient digestibility of barley, canola meal,wheat and wheat middlings, measured in vivo using rain-bow trout (Oncorhynchusmykiss). Aquaculture Nutrition8,271^277.

Cheng Z.J. & Hardy R.W. (2003a) E¡ects of extrusion pro-cessing of feed ingredients on apparent digestibility



coe⁄cients of nutrients for rainbow trout (Oncorhynchusmykiss). Aquaculture Nutrition 9,77^83.

Cheng Z.J. & Hardy R.W. (2003b) E¡ects of extrusion and ex-pelling processing, and microbial phytase supplementa-tion on apparent digestibility coe⁄cients of nutrients infull-fat soybeans for rainbow trout (Oncorhynchusmykiss).Aquaculture 218,501^514.

Choct M., Hughes R.J.,Wang J., Bedford M.R., Morgan A.J. &Annison G. (1996) Increased small intestinal fermenta-tion is partly responsible for the anti-nutritive activity ofnon-starch polysaccharides in chickens. British Journal ofPoultry Science 37,609^621.

Co¡ey M.P., Simm G., Oldham J.D., HillW.G. & BrotherstoneS. (2004) Genotype and diet e¡ects on energy balance inthe ¢rst three lactations of dairy cows. Journal of DairyScience 87, 4318^4326.

Corva P.M. & Medrano J.F. (2000) Diet e¡ects onweight gainand body composition in high growth (hg/hg) mice.Physiological Genomics 3,17^23.

Cox T.S., Shogren M.D., Sears R.G., Martin T.J. & Bolte L.C.(1989) Genetic improvement in milling and baking qual-ity of hard red winter wheat cultivars, 1919 to 1988. CropScience 29,626^631.

Dabrowski K., Rinchard J., Lee K.-J., BlomJ.H., CiereszkoA. &Ottobre J. (2000) E¡ects of diets containing gossypol onreproductive capacity of rainbow trout (Oncorhynchusmykiss). Biology of Reproduction 62, 227^234.

Davies S.J. (1985) The role of dietary ¢bre in ¢sh nutrition.In: Recent Advances in Aquaculture,Vol. 2 (ed. by J.F. Muir& R.J. Roberts), pp. 219^249.Westview Press, Boulder, CO,USA.

Dias J., Rueda-Jasso R., Panserat S., Conceicao L., Gomes E. &Dinis M. (2004) E¡ect of dietary carbohydrate-to-lipid ra-tios on growth, lipid deposition and metabolic hepatic en-zymes in juvenile Senegalese sole (Solea senegalensis,Kaup). Aquaculture Research 35,1122^1130.

DorsaW.J., Robinette H.R., Robinson E.H. & PoeW.E. (1982)E¡ects of dietary cottonseed meal and gossypol ongrowth of young channel cat¢sh. Transactions of theAmerican Fisheries Society111,651^655.

Dorsch J.A., Cook A., Young K.A., Anderson J.M., BaumanA.T.,Volkmann C.J., Murthy P.P.N. & RaboyV. (2003) Seedphosphorus and inositol phosphate of barley low phyticacid genotypes. Phytochemistry 62,691^706.

Drackley J.K. (2000) Nutritional constraints to use of soyproducts by animals. In: Soy in Animal Nutrition (ed. byJ.K. Drackley), pp. 127^170. Federated Animal Society,Savoy, IL, USA.

D’Souza N., Skonberg D.I., Camire M.E., Guthrie K.E., Mali-sonJ. & Lima L. (2005) In£uence of dietarygenistein levelson tissue genistein deposition and on the physical, chemi-cal, and sensory quality of rainbow trout, Oncorhynchusmykiss. Journal of Agriculture and Food Chemistry 53,3631^3636.

Duke J.A. (1981) Handbook of Legumes of World EconomicImportance. Plenum Press, NewYork, USA:,345pp.

Erlund I., KosonenT., Alfthan G., Maenpaa J., Perttunen K.,Kenraali J., Parantainen J. & Aro A. (2000) Pharmacoki-netics of quercetin from quercetin aglycone and rutin inhealthy volunteers. European Journal of Clinical Pharma-cology 56,545^553.

Fang N., Yu S. & Badger T.M. (2004) Comprehensivephytochemical pro¢le of soy protein isolate. Journal ofAgriculture and Food Chemistry 52, 4012^4020.

FAO. (1970) Amino acid content of foods and biological data onproteins. FAO Nutritional Studies No. 24. Food and Agri-cultural Organization of the United Nations, Rome.

FAOSTAT. (2005) Food and Agriculture Organization ofThe United Nations. http://faostat.fao.org/faostat/collections?version=ext&hasbulk=0&subset=agriculture

FAO. State of World Aquaculture. (2006) FAO Fisheries Tech-nical Paper. No.500. Rome, FAO, 145pp. http://www.fao.org/¢gis/servlet/static?dom=root&xml=aquaculture/regional_reviews_list.xml

Favier M.L., Bost P.E., Guittard C., Demigne¤ C. & Re¤ me¤ sy C.(1997) The cholesterol-lowering e¡ect of guar gum is notthe result of a simple diversion of bile acids toward fecalexcretion. Lipids 32,953^959.

Flores R.A., Eustace D.W. & Hicks K.B. (2004) Experimentalmilling of barley samples. Meeting Abstract 2004. Ameri-can Association of Cereal Chemists, San Diego, CA, USA,19^22 September. Poster 243.

Floreto E.A.T., Baker R.C. & Brown P.B. (2000) The e¡ects ofsoybean-based diets, with and without amino acid sup-plementation, on growth and biochemical compositionof juvenile American lobster, Homarus americanus.Aquaculture189, 211^235.

Formica J.V. & Regelson W. (1995) Review of the biology ofquercetin and related bio£avonoids. Food and ChemicalToxicology 33,1061^1080.

Fowler L.G. (1980) Substitution of soybean and cottonseedproducts for ¢sh meal in diets fed to chinook and cohosalmon.The Progressive Fish-Culturist 42,87^91.

Francis G., Makkar H.P.S. & Becker K. (2001) Antinutritionalfactors present in plant-derived alternate ¢sh feed ingre-dients and their e¡ects in ¢sh. Aquaculture199,197^227.

Francis G., Makkar H.P.S. & Becker K. (2005) Quillaja sapo-nins ^ a natural growth promoter for ¢sh. Animal FeedScienceTechnolology121,147^157.

Furuya W.M., Pezzato L.E., Barros M.M., Pezzato A.C.,FuruyaV.R.B. & Miranda E.C. (2004) Use of ideal proteinconcept for precision formulation of amino acid levels in¢sh-meal free diets for juvenile Nile tilapia (Oreochromisniloticus L.). Aquaculture Research 35,1110^1116.

Garcia-Abiado M.A., Mbahinzireki G., Rinchard J., Lee K.J. &Dabrowski K. (2005) E¡ect of diets containing gossypol onblood parameters and spleen structure in tilapia, Oreo-chromic sp., reared in a recirculating system. Journal ofFish Diseases 27,359^368.

Gibbon B.C. & Larkins B.A. (2005) Molecular geneticapproaches to developing quality protein maize. Trendsin Genetics 21, 227^233.



Glencross B.D., Boujardand T. & Kaushick S.J. (2003) In£u-ence of oligosaccharides on the digestibility of lupinmeals when fed to rainbow trout, Oncorhynchus mykiss.Aquaculture 219,703^713.

Glencross B., Curnow J., HawkinsW., Kissil G.W. & PetersonD. (2003) Evaluation of the feed value of a transgenicstrain of the narrow-leaf lupin (Lupinus angustifolius) inthe diet of the marine ¢sh, Pagrus auratus. AquacultureNutrition 9,197^206.

Glencross B.D., Carter C.G., Duijster N., Evans D.R., Dods K.,McCa¡erty P., Hawkins W.E., Maasand R. & Sipsas S.(2004) A comparison of the digestibility of a range of lu-pin and soybean protein products when fed to eitherAtlantic salmon (Salmo salar) or rainbow trout (Oncor-hynchus mykiss). Aquaculture 237,333^346.

Glencross B., Evans D., HawkinsW. & Jones B. (2004) Evalua-tion of dietary inclusion of yellow lupin (Lupinus luteus)kernel meal on the growth, feed utilisation and tissuehistology of rainbow trout (Oncorhynchus mykiss). Aqua-culture 235, 411^422.

Graybosch R.A., Souza E. & BerzonskyW. (2003) Registra-tion of 19 waxy spring wheat germplasm. Crop Science43,1491^1492.

Grigorakis K.,Taylor K.D.A. & Alexis M.N. (2001) Organolep-tic and volatile aroma compounds comparison of wildand cultured gilthead sea bream (Sparus aurata): sensorydi¡erences and possible chemical basis. Aquaculture 225,109^119.

Guillon F. & Champ M.M.J. (2002) Carbohydrate fractions oflegumes: uses in human nutrition and potential forhealth. BritishJournal of Nutrition 88, S293^S306.

Guttieri M.J., Bowen D., Dorsch J.A., Souza E. & Raboy V.(2004) Identi¢cation and characterization of a low phyticacid wheat. Crop Science 44, 418^424.

Hardy R.W. & Barrows F.T. (2000) Diet formulationandman-ufacturing. In: Fish Nutrition (ed. by J.E. Halver & R.W.Hardy), 3rd edn, pp. 505^600. Academic Press, NewYork,NY, USA.

Herman E.M. (2004) Allergenic responses to legume pro-teins. In: Genomics for Legume Crops (ed. by C. Brummer& R.Wilson), pp.326^347. American Oil Chemists SocietyPress, Champaign, IL, USA.

Herman R.L. (1970) E¡ects of gossypol on rainbow trout(Salmo gairdneri). Journal of Fish Biology 2, 293^297.

Hickling D. (2003) Canadian Feed Peas, Industry Guide. PulseCanada,Winipeg, MB, Canada,35pp.

Hicks K.B.,Taylor F., Kohout K.M., Kurantz M.J.,Thomas J.L.,O’Brien D.J., Johnston D., Flores R.A., Moreau R.A. & HootL. (2004) Developing barley as amajor feedstock for fuel etha-nol production. Meeting Abstract 2004. International FuelEthanolWorkshop & Trade Show, 22^25 June, Session 7,Feedstocks for Ethanol Production, Madison,WI, USA.

Hostetler H., Collodi P., Devlin R. &MuirW. (2005) Improvedphytate phosphorus utilization by Japanese medakatransgenic for theAspergillus niger phytase gene. Zebra¢sh2,19^31.

HouY.C., Chao P.D., Ho H.J.,Wen C.C. & Hsiu S.L. (2003) Pro-found di¡erence in pharmacokinetics betweenmorin andits isomer quercetin in rats. Journal of Pharmacology 55,199^203.

van den Ingh T.S.G.A.M., Krogdahl —., Olli J.J., HendrixH.G.C.J.M. & Koninkx J.G.J.F. (1991) E¡ects of soybean con-taining diets on the proximal and distal intestine inAtlantic salmon (Salmo salar): a morphological study.Aquaculture 94, 297^305.

van den Ingh T.S.G.A.M., Olli J.J. & Krogdahl —. (1996)Alcohol-soluble components in soybeans cause morpho-logical changes in the distal intestine of the Atlanticsalmon, Salmo salar L. Journal of Fish Disease 19,47^53.

Jenkins P.G., Harris J.E. & Pulsford A.L. (1991) Enhanced en-teric uptake of humangammaglobulin by Quil-A saponinin Oreochromis mossambicus. Fish and Shell¢sh Immunol-ogy1, 279^295.

Jenkins D.J., Kendall C.W., Vuksan V., Vidgen E., Parker T.,Faulkner D., Mehling C.C., Garsetti M., Testolin G.,Cunnane S.C., RyanM.A. & Corey P.N. (2002) Soluble ¢berintake at a dose approved by the US food and drug admin-istration for a claim of health bene¢ts: serum lipid riskfactors for cardiovascular disease assessed in a rando-mized controlled crossover trial. American Journal ofClinical Nutrition 75,834^839.

Johnsen P.B. & Dupree H.K. (1991) In£uence of feed ingredi-ents on the £avor quality of farm-raised cat¢sh. Aquacul-ture 96,139^150.

Jones E., ZaczekV., MacLeod M. & Hocking P.M. (2004) Gen-otype, dietary manipulation and food allocations a¡ectindices of welfare in broiler breeders. British Journal ofPoultry Science 45,725^737.

Kaiser C.R., Bowman J.G.P., Surber L.M.M., BlakeT.K. & Bor-owski J.J. (2004) Variation in apparent component digest-ibility of barley in the rat from the core collection of theUSDA nation small grains collection. Animal Feed ScienceTechnology113,97^112.

Kaput J., Klein K.G., Reyes E.J., KibbeW.A., Cooney C.A., Jo-vanovic B.,VisekW.J. & Wol¡ G.L. (2004) Identi¢cation ofgenes contributing to the obese yellow Avy phenotype:caloric restriction, genotype, diet � genotype interac-tions. Physiological Genomics18,316^324.

Kasumyan A.O. & Doving K.B. (2003) Taste preferences in¢shes. Fish and Fisheries 4, 289^347.

Keembiyehetty C.N. & Gatlin D.M. III. (1997) Performance ofsunshine bass fed soybean-meal-based diets supplemen-ted with di¡erent methionine compounds. The Progressi-veFish-Culturist 59, 25^30.

Kim Y.S., Flores R.A., Chung O.K. & Bechtel D.B. (2003)Physical and chemical characterization of wheat £ourmilling co-products. Journal of Food Process Engineering26, 469^488.

Krogdahl —. (1996) E¡ects of soybean antinutrients on in-testinal environment, lipid hydrolysis and utilization onAtlantic salmon (Salmo salar, L). I e¡ects of antinutrient



on the nutritional value of legume diets. (ed. byS. Bardocs, F.V. Nekrep & A. Pusztai),Vol.3 (pp.74^81.

Krogdahl —. & Holm H. (1983) Pancreatic proteinases fromman, trout, rat, pig, cow, chicken, mink and fox. Enzymeactivities and inhibition by soybeanand lima bean protei-nase inhibitors. Comparative Biochemistry and Physiology74B, 403^409.

Krogdahl —., Bakke-McKellep A.M. & Baeverfjord G. (2003)E¡ects of graded levels of standard soya bean meal onintestinal structure, mucosal enzyme activities, andpancreatic response in Atlantic salmon (Salmo salar L).Aquaculture Nutrition 9,361^371.

Kuz’minaV.V. (1996) In£uence of age on digestive enzyme ac-tivity in some freshwater teleosts.Aquaculture148,25^37.

Latonnelle K., FostierA., LeMenn F. & Bennetau-Pelissero C.(2002) Binding a⁄nities of hepatic nuclear estrogen re-ceptors for phytoestrogens in rainbow trout (Oncor-hynchus mykiss) and Siberian sturgeon (Acipenser baeri).General and Comparative Endocrinology129,69^79.

Latonnelle K., LeMenn F., Kaushik S.J. & Bennetau-PelisseroC. (2002) E¡ects of dietary phytoestrogens in vivo and invitro in rainbow trout and Siberian sturgeon: interestsand limits of the in vitro studies of interspecies di¡er-ences. General and Comparative Endocrinology126,39^51.

Laurie C.C., Chasalow S.D., LeDeauxJ.R., McCarroll R., BushD., Hauge B., Lai C., Clark D., Rocheford T.R. & Dudley J.W.(2004) The genetic architecture of response to long-termarti¢cial selection for oil concentration in the maizekernel. Genetics168, 2141^2155.

Lee K.J., Dabrowski K., Blom J.H., Bai S.C. & Stromberg P.C.(2002) A mixture of cottonseed meal, soybean meal andanimal byproduct mixture as a ¢sh meal substitute:growth and tissue gossypol enantiomers in juvenilerainbow trout (Oncorhynchus mykiss). Journal of AnimalPhysiology and Nutrition 86, 201^213.

Lee K.J., Rinchard J., Dabrowski K., Babiak I., Ottobre J.S. &Christensen J.E. (2006) Long-term e¡ects of dietary cot-tonseed meal on growth and reproductive performanceof rainbow trout: three-year study. Animal Feed ScienceandTechnology126,93^106.

Levrat M.A., Moundras C.,YounesH.,MorandC., Demigne C.& Re¤ me¤ sy C. (1996) E¡ectiveness of resistant starch, com-pared to guar gum, in depressing plasma cholesterol andenhancing fecal steroid excretion. Lipids 31,1069^1075.

Liener I.E. (2000) Non-nutritive factors and bioactive com-pounds in soy. In: Soy in Animal Nutrition (ed. by J.K.Drackley), pp.127^170. Federated Animal Science Society,Savoy, IL, USA.

Linberg J.E., Arvidsson A. & Wang J. (2003) In£uence ofnaked barley cultivar with normal, amylose-rich or amy-lopectin-rich starch and enzyme supplementation on di-gestibility and piglet performance. Animal Feed ScienceandTechnology104,121^131.

Luckett D. (2004) Lupin bean: a bitter contamination risk forsweet albus lupins. NSW Department of Primary Indus-tries, Agnote DPI-498,18 August 2004 (1st edn).

Lupton J.R., Robinson M.C. & Morin J.L. (1994) Cholesterol-lowering e¡ect of barley bran £our and oil. Journal of theAmerican Dietetic Association 94,65^70.

Mambrini M., RoemA.J., Cravedi J.P., Lalles J.P. & Kaushik S.J.(1999) E¡ects of replacing ¢sh meal with soy proteinconcentrate and of DL-methionine supplementation inhigh-energy, extruded diets on the growth and nutrientutilizationof rainbow trout,Oncorhynchusmykiss. Journalof Animal Science 77, 2990^2999.

Manach C., Morand C., Crespy V., Demigne C., Texier O.,Regerat F. & Remesy C. (1998) Quercetin is recoveredin human plasma as conjugated derivatives whichretain antioxidant properties. FEBS Letter 426, 331^336.

Manach C., Texier O., Morand C., Crespy V., Regerat F., De-migne C. & Remesy C. (1999) Comparison of the bioavail-ability of quercetin and catechin in rats. Free Radicals inBiology and Medicine 27,1259^1266.

Martin J.H., LeonardW.H. & Stamp D.L. (1976) Principles ofField Crop Production, 3rd edn. Macmillan Publishing,NewYork, USA.

Martinez Nieto L., Garrido Hoyos S.E., Camacho Rubio F.,Garcia Pareja M.P. & Ramos Cormenzana A. (1993) Thebiological puri¢cation of waste products from olive oil ex-traction. BioresourceTechnology 43, 215^219.

Milligan S.R., KhanO. &NashM. (1998) Competitive bindingof xenobiotic oestrogens to rat alpha-fetoprotein and tosex steroid binding proteins in human and rainbow trout(Oncorhynchus mykiss) plasma. General and ComparativeEndocrinology112,89^95.

Morell M.K. & Myers A.M. (2005) Towards the rationaldesign of cereal starches. Current Opinion in Plant Biology8, 204^210.

Moundras C., Behr S.R., Re¤ me¤ sy C. & Demigne¤ C. (1997)Fecal losses of sterols and bile acids induced by feedingrats guar gum are due to greater pool size and liverbile acid secretion. Journal of Nutrition 127, 1068^1076.

Mukhopadhyay N. & RayA.K. (1999) E¡ect of fermentationon the nutritive value of sesame seed meal in the dietsfor rohu, Labeo rohita (Hamilton), ¢ngerlings. AquacultureNutrition 5, 229^236.

Nelson P. & Delane R. (1990) Producing lupins in WesternAustralia. In: Bulletin 4176.WA Department of Agriculture.(ed. by E. Lawson). South Perth,Western Australia.

NgW., Lim H., Lim S. & Ibrahim C. (2002) Nutritive value ofpalm kernel meal pretreated with enzyme or fermentedwithTrichoderma kongii (Oudemans) as a dietary ingredi-ent for red hybrid tilapia (Oreochromis sp.). AquacultureResearch 33,1199^1207.

NRC. (1993) Nutrient Requirements of Fish. Natioal ResearchCouncil. National Academy Press,Washington, DC, USA,114pp.

NRC. (1994) Nutrient Requirements of Poultry. NationalResearch Council. National Academy Press, Washington,DC, USA,155pp.



NRC. (1998) Nutrient Requirements of Swine. NationalResearch Council. National Academy Press, Washington,DC, USA,189pp.

Oda K., Matsuda H., Murakami T., Katayama S., Ohgitani T.& YoshikawaM. (2003) Relationship betweenadjuvant ac-tivity and amphipathic structure of soyasaponins.Vaccine21, 2145^2151.

Okot-KotberM.,Yong K.J., Bagorogoza K. & Liavoga A. (2003)Phytase activity in extracts of £our and bran fromwheatcultivars: enhanced extractability with [b]-glucanaseand endo-xylanase. Journal of Cereal Science 38,307^315.

Olli J.J. & Krogdahl —. (1994) Nutritive value of four soybeanproducts as protein sources in diets for rainbow trout (On-corhynhcus mykiss,Walbaum) reared in fresh water. ActaAgriculturae Scandinavica 44,185^192.

Olli J.J., Hjelmeland K. & Krogdahl —. (1994) Soybean trypsininhibitors in diets for Atlantic salmon (Salmo salar, L):e¡ects on nutrient digestibilities and trypsin in pyloriccaeca homogenate and intestinal content. ComparativeBiochemistry and Physiology109A,923^928.

Olli J.J., Krogdahl —., van dan InghT.S.G.A.M. & Bratt�s L.E.(1994) Nutritive value of four soybean products in dietsfor Atlantic salmon (Salmo salar, L.). Acta AgriculturaeScandinavica Section A, Animal Science 44,50^60.

Olsen R.E., Myklebust R., Kryvi H., MayhewT.M. & Ringo E.(2001) Damaging e¡ect of dietary inulin on intestinalenterocytes in Arctic charr (Salvelinus alpinus L.). Aqua-culture Research 32,931^934.

Overturf K., Raboy V., Cheng Z.J. & Hardy R.W. (2003)Mineral availability from barley low phytic acid grains inrainbow trout (Oncorhynchus mykiss) diets. AquacultureNutrition 9, 239^246.

Overturf K., Bullock D., LaPatra S. & Hardy R. (2004) Genet-ic selection and molecular analysis of domesticatedrainbow trout for enhanced growth on alternative dietsources. Environmental Biology of Fishes 69, 409^418.

Pelissero C., LeMenn F. & Kaushick S. (1991) Estrogenic e¡ectof dietary soya bean meal on vitellogenesis in cultured Si-berian sturgeon Acipenser baeri. General and ComparativeEndocrinology 83, 447^457.

Pelissero C., Bennetau B., Babin P., LeMenn F. & Dunogues J.(1991) The estrogenic activity of certain phytoestrogensin the Siberian sturgeon Acipenser baeri. Journal of SteroidBiochemistry andMolecular Biology 38, 293^299.

Pezzato L.E., deOliveira A.C.B., Dias E., Barros M.M. & Pezza-to A.C. (2004) Weight gain and anatomopathological dis-turbs on Nile tilapia fed with cocoa meal. PesquisaAgropecuaria Brasileira 31,375^378.

Piskula M.K. & Terao J. (1998) Accumulation of (� )-epicate-chin metabolites in rat plasma after oral administrationand distribution of conjugation enzymes in rat tissues.Journal of Nutrition128,1172^1178.

Plakas S.M., Lee T.-C. & Wolke R.E. (1985) Absence of overttoxicity from feeding the £avonol, quercetin, to rainbowtrout (Salmo gairdneri). Food and Chemical Toxicology 23,1077^1080.

Pollack S.J., Ottinger M.A., Sullivan C.V. & Woods L.C. III.(2003) The e¡ects of the soy iso£avone genistein on thereproductive development of striped bass.North AmericanJournal of Aquaculture 65, 226^234.

PomeranzY., Carvajal M.I., Hoseney R.C. &Ward A.B. (1970)Wheat germ in bread making. I. Composition of germlipids and germ protein fractions. Cereal Chemistry 47,373^380.

RaboyV. & DickinsonD.B. (1993) Phytic acid levels in seeds ofGlycine max and G. soja as in£uenced by phosphorusstatus. Crop Science 33,1300^1305.

Refstie S., StorebakkenT. & Roem A.J. (1998) Feed consump-tion and conversion in Atlantic salmon (Salmo salar) feddiets with ¢sh meal, extracted soybean meal or soybeanmeal with reduced content of oligosaccharides, trypsininhibitors, lectins and soy antigens. Aquaculture 162,301^312.

Refstie S., Svihus B., Shearer K.D. & Storebakken T. (1999)Nutrient digestibility in Atlantic salmon and broilerchickens related to viscosity and non-starch polysacchar-ide content in di¡erent soyabean products. Animal FeedScience andTechnology 79,331^345.

Refstie S., Sahlstrom S., Brathen E., Baeverfjord G. & Kroge-dal P. (2005) Lactic acid fermentation eliminates indiges-tible carbohydrates and antinutritional factors insoybean meal for Atlantic salmon (Salmo salar). Aquacul-ture 246,331^345.

Riche M. & Brown P.B. (1996) Availability of phosphorusfrom feedstu¡s fed to rainbow troutOncorhynchusmykiss.Aquaculture142, 269^282.

Rinchard J., Mbahinzireki G., Dabrowski K., Lee K.-J.,Garcia-Abiado M.A. & Ottobre J.S. (2002) E¡ects of dietarycottonseed meal protein level on growth, gonad develop-ment and plasma sex steroid hormones of tropical ¢shtilapia,Oreochromis sp.Aquaculture International10,11^28.

Rinchard J., Lee K.J., Czesny S., CiereszkoA. & Dabrowski K.(2003) E¡ect of feeding cottonseed meal-containing dietsto broodstock rainbow trout and their impact on thegrowth of their progenies. Aquaculture 227,77^87.

Robinson E.H. & Li M.H. (1994) Use of plant proteins in cat-¢sh feeds: replacement of soybean meal with cottonseedmeal and replacement of ¢shmealwith soybeanmeal andcottonseed meal. Journal of theWorld Aquaculture Society25, 271^276.

Robinson E.H. & TierschT.R. (1995) E¡ects of long-term feed-ing of cottonseed meal on growth, testis development,and spermmotility of male channel cat¢sh Ictalurus punc-tatus brood¢sh. Journal of the World Aquaculture Society26, 426^431.

Roche J.R., Kolver E.S., de Veth M.J. & Napper A. (2001) Dietand genotype a¡ect plasma calcium, magnesium andphosphorus concentrations in the periparturient cow.Proceedings of the New Zealand Society of Animal Produc-tion 61,168^171.

Rohlo¡ J. (2003) Cultivation of Herbs and Medicinal Plants inNorway ^ Essential Oil Production and Quality Control. The



Plant Biocentre, Department of Biology, Faculty of Natur-al Sciences and Technology, Norwegian University ofScience and Technology (NTNU), Trondheim, Norway89pp, ISBN 82-471-5177-4.

Rohlo¡ J. (2004) Essential oil drugs ^ terpene composition ofaromatic herbs. In: Production Practices and QualityAssessment of Food Crops. Vol. 3: Quality Handling andEvaluation (ed. by R. Dris & S.M. Jain), pp.73^128. KluwerAcademic Publishers, Dordrecht, the Netherlands.

Roy P.K., Munshi J.D. & Dutta H.M. (1990) E¡ect of saponinextracts on morpho-histology and respiratory physiologyof an air-breathing ¢sh, Heteropneustes Fossilis (Bloch).Journal of Freshwater Biology 2,135^145.

Russett C. (2002) Soy Protein Concentrates forAnimal Feeds.Animal Nutrition Research Notes.The Solae. SPC-T-47, FortWayne, IN,12 pp.

SealeyW.M. & Gatlin D.M. III (1999) Overviewof nutritionalstrategies a¡ecting health of marine ¢sh. Journal ofApplied Aquaculture 9,11^26.

Shahidi F. & Naczk M. (1995) Food Phenolics.Technomic Pub-lishing Co. Lancaster, PA,331pp.

Sklan D., Prag T. & Lupatsch I. (2004) Apparent digestibilitycoe⁄cients of feed ingredients and their prediction indiets for tilapia Oreochromis niloticus � Oreochromis aur-eus (Teleostei, Cichlidae). Aquaculture Research 35,358^364.

Skrede G., StorebakkenT., SkredeA., Sahlstrom S., SorensenM., Shearer K.D. & Slinde E. (2002) Lactic acid fermenta-tion of wheat and barley whole meal £ours improves di-gestibility of nutrients and energy in Atlantic Salmon(Salmo salar L.) diets. Aquaculture 210,305^321.

Souza E.,Tyler J.M., Kephart K.D. & Kruk M. (1993) Geneticimprovement in milling and baking quality of hard redspring wheat cultivars. Cereal Chemistry 70, 280^285.

StepanskyA.,YaoY.,Tang G. & Galili G. (2004) Regulation oflysine catabolism in Arabidopsis through concertedlyregulated synthesis of the two distinct gene products ofthe composite ATLKR/SDH locus. Journal of ExperimentalBotany 56,525^536.

Stickney R.R. & Shumway S.E. (1974) Occurrence of cellu-lase activity in the stomachs of ¢shes. Journal of FishBiology 6,779^790.

Storebakken T. (1985) Binders in ¢sh feeds I. E¡ect of algi-nate and guar gum on growth, digestibility, feed intakeand passage through the gastrointestinal tract of rainbowtrout. Aquaculture 47,11^26.

Storebakken T. & Austreng E. (1987) Binders in ¢sh feedsII. E¡ect of di¡erent alginates on the digestibility ofmacronutrients in rainbow trout. Aquaculture 60,121^131.

StorebakkenT., Kvien I.S., Shearer K.D., Grisdale-Helland B.,Helland S.J. & Berge G.M. (1998) The apparent digestibilityof diets containing ¢sh meal, soybean meal or bacterialmeal fed to Atlantic salmon (Salmo salar): evaluation ofdi¡erent faecal collection methods. Aquaculture 169,195^210.

StorebakkenT., Refstie S. & Ruyter B. (2000) Soy products asfat and protein sources in ¢sh feeds for intensive aqua-culture. In: Soy in Animal Nutrition (ed. by J.K. Drackley),pp. 127^170. Federated Animal Science Society, Savoy,IL, USA.

Sugiura S.H., Dong F.M., Rathbone K.C. & Hardy R.W. (1998)Apparent protein digestibility and mineral availabilitiesin various feed ingredients for salmonid feeds. Aquacul-ture159,177^202.

Sugiura S.H., RaboyV.,Young K.A., Dong F.M. & Hardy R.W.(1999) Availability of phosphorus and trace elements inlow-phytate varieties of barleyand corn for rainbow trout(Oncorhynchus mykiss). Aquaculture170, 285^296.

Takagi S., Shimeno S., Hosokawa H. & Ukawa M. (2001)E¡ect of lysine and methionine supplementation to asoy protein concentrate diet for red sea bream Pagrusmajor. Fisheries Science 67,1088^1096.

Takeuchi T. (2001) A review of feed development for earlylife stages of marine ¢n¢sh in Japan. Aquaculture 200,203^222.

Thiessen D.L., Campbell G.L. & Adelizi P.D. (2003) Digestibil-ity and growth performance of juvenile rainbow trout(Oncorhynchus mykiss) fed with pea and canola products.Aquaculture Nutrition 9,67^75.

Thiessen D.L., Campbell G.L. & Tyler R.T. (2003) Utilizationof thin distillers solubles as a palatability enhancer inrainbow trout (Oncorhynchus mykiss) diets containingcanola meal or air-classi¢ed pea protein. AquacultureNutrition 9,1^10.

Tollefsen K.E., Ovrevik J. & Stenersen J. (2004) Binding ofxenoestrogens to the sex steroid-¢nding protein inplasma from Arctic charr (Salvelinus alpinus L.). Compara-tive Biochemistry Physiolology139 (C),127^133.

Twibell R.G. &Wilson R.P. (2004) Preliminary evidence thatcholesterol improves growth and feed intake of soybeanmeal-based diets in aquaria studies with juvenile channelcat¢sh, Ictalurus punctatus. Aquaculture 236,539^546.

Tzchori I., Degani G., Elisha R., Eliyahu R., Hurvitz A.,VayaJ.& Moav B. (2004) The in£uence of phytoestrogens andoestradiol-17b on growth and sex determination in theEuropean eel (Anguilla anguilla). Aquaculture Research 35,1213^1219.

Ulloa J.B., vanWeerd J.H., Huisman E.A. & Verret J.A.J. (2004)Tropical agricultural residues and their potential uses in¢sh feeds: the Costa Rican situation.Waste Management24,87^97.

USDA-National Agricultural Statistics Service. (2005)http://www.nass.usda.gov.

Venou B., Alexis M.N., Fountoulaki E., Nengas I., Apostolo-poulou M. & Castritsi-Cathariou I. (2003) E¡ect of extru-sion of wheat and corn on gilthead sea bream (Sparusaurata) growth, nutrient utilization e⁄ciency, rates ofgastric evacuation and digestive enzyme activities.Aquaculture 225, 207^223.

Vivas M., Sanchez-Vazquez F.J., Garcia G. & Madrid J.A.(2003) Macronutrient self-selection in European sea bass



in response to dietary protein or fat restriction. Aquacul-ture Research 34, 271^280.

Wahstrom A.,Tauson R. & Elwinger K. (1998) E¡ects on plu-mage condition, health and mortality of dietary oats/wheat ratios to three hybrids of laying hens in di¡erenthousing systems. Acta Agriculturae Scandinavica A,Animal Science 48, 250^259.

Waldroup P.W. & Smith K.J. (1989) Animal feed uses oflegumes. In: Legumes Chemistry, Technology and HumanNutrition (ed. by R.H. Matthews), Marcel Dekker, NewYork.

Weber L.P., KiparissisY., Hwang G.S., Niimi A.J., Janz D.M. &Metcalfe C.D. (2002) Increased cellular apoptosis afterchronic aqueous exposure to nonylphenol and quercetinin adult medaka (Oryzias latipes). Comparative Biochemis-try and Physiology Part C: Toxicology and Pharmacology131,51^59.

Wilkes G.J., Conte E.D.,Yonkyoung K., Holcomb M., Suther-land J.B. & Miller D.W. (2000) Sample preparation for the

analysis of £avours and o¡-£avours in foods. Journal ofChromatography 880A,3^33.

Wilson R.P. (2002) Amino acids and proteins. In: Fish Nutri-tion (ed. by J.E. Halver & R.W. Hardy),3rd edn, pp.144^181.Academic Press, NewYork, NY, USA.

Wittig J., Herderich M., Graefe E.U. & Veit M. (2001) Identi¢-cation of quercetin glucuronides in human plasma byhigh-performance liquid chromatography^tandem massspectrometry. Journal of Chromatography B: BiomedicalSciences and Applications 753, 237^243.

Yamamori M., Fujita S., Hayakawa K. &Matsuki J. (2000) Ge-netic elimination of a starch granule protein, SGP-1, ofwheat generates an altered starch with apparent highamylose.Theoretical and Applied Genetics101, 21^29.

YamamotoT., ShimaT., Furuita H. & Suzuki N. (2002) In£u-ence of feeding diets with and without ¢sh meal by handand by self-feeders on feed intake, growth and nutrientutilization of juvenile rainbow trout (Oncorhynchus my-kiss). Aquaculture 214, 289^305.



Expanding the utilization of sustainable plant products in ... · compounded rate of 9.2% per year...

Documents

Transcript of Expanding the utilization of sustainable plant products in ... · compounded rate of 9.2% per year...