Carotenoid and lipid content in muscle of Atlantic salmon, Salmo salar, transferred to seawater

Comparative Biochemistry and Physiology Part B 138(2004)29–401096-4959/04/$-see front matter ᮊ2004Elsevier Inc.All rights reserved.doi:10.1016/j.cbpc.2004.01.011Carotenoid and lipid content in muscle of Atlantic salmon,Salmosalar ,transferred to seawater as 0q or 1q smolts ૾T.Ytrestøyl ,G.Coral-Hinostroza ,B.Hatlen ,D.H.F .Robb ,B.Bjerkeng *a a ,1b ,2b a ,AKVAFORSK (Institute of Aquaculture Research AS ),N-6600Sunndalsøra,NorwayaEWOS Innovation AS,N-4335Dirdal,Norwayb Received 9October 2003;received in revised form 19January 2004;accepted 21February 2004AbstractAccumulation of lipids and carotenoids,including 49-hydroxyechinenone (49-hydroxy-b ,b -carotene-4-one ),growth and condition factor were investigated in Atlantic salmon (Salmo salar )transferred to seawater as 0q and 1q smolts.Salmon were fed a diet with 30mg y kg astaxanthin (3,39-dihydroxy-b ,b -carotene-4,49-dione )and 30mg y kg canthaxanthin (b ,b -carotene-4,49-dione )for 35weeks.The 0q smolt contained more carotenoids than the 1q smolt when mass differences were corrected for (P -0.0001),a difference also reflected by the tristimulus colour measurements (C1E a*-and b*-values ).Astaxanthin and canthaxanthin comprised more than 93%of the total carotenoids,but small differences were observed in carotenoid composition.The condition factor was significantly higher in 0q than 1q smolts after correction for mass differences (P -0.01).There was a high correlation between ln-transformed muscle lipid (%)and ln-transformed body mass for 0q (R s 0.94)and 1q smolts (R s 0.97).The canthaxanthin metabolite 49-hydroxyechi-22nenone was isolated from muscle of Atlantic salmon fed a diet supplemented with canthaxanthin.It was characterised and identified by its absorption maximum (l s 458nm in n -hexane ),mass spectrometry (M s 566)and co-q max chromatography with authentic standard obtained by NaBH -reduction of canthaxanthin on thin-layer chromatography 4and HPLC.HPLC of the camphanates of 49-hydroxyechinenone revealed a stereoselective transformation in favour of the (49S )-isomer,the (49S )and (49R )-isomers comprising approximately 81and 19%of the total 49-hydroxyechinenone,respectively.The percentage of 49-hydroxyechineone of total carotenoids ranged from 1.3to 3.1%and declined with fish size (P -0.001).We conclude that effects of time of seawater transfer of Atlantic salmon smolts have significant effect on carotenoid accumulation and other quality traits.The detailed biochemical and physiological basis for these differences require further elucidation.ᮊ2004Elsevier Inc.All rights reserved.Keywords:Atlantic salmon;Canthaxanthin;49-Hydroxyechinenone;Pigmentation;Colour;Condition factor;Fat;Quality૾Preliminary results from this study were presented at the 10th International Symposium on Nutrition and Feeding in Fish,Rhodes,Greece,2–7June,2002,Abstract Book,p.228.*Corresponding author.Tel.:q 47-71-69-53-05;fax:q 47-71-69-53-01.E-mail address:bjorn.bjerkeng@akvaforsk.nlh.no (B.Bjerkeng ).Present address:Instituto Nacional de Ciencias Medicas y de Nutricion ‘Salvador Zubiran’,Departamento de NutricionAnimal,1´´´´Vasco de Quiroga 15,C.P .14000,Mexico City,Mexico.´Present address:AKV AFORSK (Institute of Aquaculture Research AS ),N-6600Sunndalsora,Norway.230T.Ytrestøyl et al./Comparative Biochemistry and Physiology Part B138(2004)29–401.IntroductionIn the wild,Atlantic salmon(Salmo salar)migrate towards the sea during spring and the ageat smolting ranges from1q to5q depending ontemperature and natural photoperiod at a givenlatitude(Metcalfe and Thorpe,1990).Smolting isfundamentally related to development,and juvenilesalmon must be of an appropriate size-relateddevelopmental stage to be capable to smoltify(McCormick et al.,1998).To enable stable and efficient production of smolts throughout the yearit is current practice to advance smolting to autumnat age0q.In intensive salmon farming,food isabundant and advanced smolting is achieved bythermal and photoperiodic control(Gaignon andQuemener,1992).The photoperiod to which salm-on parr are exposed early in life may causeprecocious maturation(Duston and Saunders,1995;Berrill et al.,2003),and different growthprofiles are exhibited which depend on the differ-ent photoperiods and temperatures that are expe-rienced by fish before transfer to the sea(Duncanet al.,2002).Nonetheless,growth rates over theentire production cycle may be similar(Duncan etal.,1998).The knowledge on effects of smoltproduction practice on important parameters suchas carotenoid and lipid accumulation in the muscleis sparse.However,Bjerkeng et al.(1992)foundan increased rate of deposition of carotenoids inthe muscle of rainbow trout(Oncorhynchusmykiss)following smoltification and seawater transfer.A primary objective of the present study was,therefore,to compare carotenoid deposition in muscle of Atlantic salmon transferred to sea as 0q and1q smolts during a35weeks production period.Fillet colour is an important quality parameterfor salmonid fishes,and pigment feeding is regard-ed as the most important management practice formarketing of farmed salmon(Moe,1990).Astax-anthin(3,39-dihydroxy-b,b-carotene-4,49-dione)isthe predominant carotenoid of wild Atlantic salm-on(Khare et al.,1973;Schiedt et al.,1981,1986)and astaxanthin and canthaxanthin(b,b-carotene-4,49-dione),either alone or in combination are thecarotenoids commonly used for pigmentation ofsalmonid fishes in aquaculture(cf.Shahidi et al.,1998).Carotenoid feeding is expensive,albeitindispensible for Atlantic salmon farming,and thecost of supplements may amount to approximately15%of total feed costs.Whereas astaxanthin is more efficiently accumulated in the muscle of rainbow trout than canthaxanthin(reviewed by Storebakken and No1992),the opposite appears to be true for Atlantic salmon(Buttle et al.,2001; Baker et al.,2002).Thus,a higher level of can-thaxanthin than astaxanthin is absorbed into blood when Atlantic salmon are given the same dietary inclusion levels(Kiessling et al.,2003).These two carotenoids are biochemically active and meta-bolised reductively in salmonid fishes and serve as precursors of vitamin A(Schiedt et al.,1985). The ability to metabolise and accumulate carote-noids in muscle and skin changes with age and physiological status(Storebakken and No1992). Before the smolt stage carotenoids are deposited mainly in the skin(Storebakken et al.,1987; Bjerkeng et al.,1992).As the fish grow the ability to deposit carotenoids in the muscle increases and the concentration in the flesh increases while the concentration in the skin is reduced(Bjerkeng et al.,1992).Astaxanthin is deposited in the free form in the muscle while most of the carotenols present in the skin are esterified(cf.Schiedt et al., 1988a,b).As the fish enter sexual maturation carotenoids are redistributed from the flesh to the skin and gonads,or transformed into colourless compounds(Bjerkeng et al.,1992),a process possibly triggered by sex hormones(Bjerkeng et al.,1999).Idoxanthin(3,39,49-trihydroxy-b,b-car-otene-4-one),a metabolite of astaxanthin,may accumulate in the muscle when astaxanthin is fed to Atlantic salmon(Schiedt et al.,1988a,b,1989) and Arctic charr(Salvelinus alpinus;Aas et al., 1997;Hatlen et al.,1997;Bjerkeng et al.,2000). The proportion of idoxanthin tends to decrease with fish size and age.The enzymatic reduction of astaxanthin to idoxanthin is stereoselective and leads to formation of the(49R)isomer of idoxan-thin irrespective of the configuration at C(39)in astaxanthin(Schiedt et al.,1988a).Analogously, 49-hydroxyechineone(49-hydroxy-b,b-carotene-4-one,cf.Fig.1for structures)is a metabolite of canthaxanthin previously isolated from the skin of Atlantic salmon(Schiedt et al.,1988b)and rain-bow trout(Bjerkeng et al.,1992)and from liver and egg yolk in laying hens(Schiedt,1989)fed diets containing canthaxanthin.The reduction of canthaxanthin is also stereoselective favouring the formation of(49S)-hydroxyechineone(Schiedt, 1989;Bjerkeng et al.,1992).Recently,we found a component tentatively identified as49-hydroxyechineone by high-per-31T.Ytrestøyl et al./Comparative Biochemistry and Physiology Part B 138(2004)29–40Fig.1.Reductive transformation of canthaxanthin to (49R )-and (49S )-hydroxyechinenone.Table 1Composition of feeds used Feed compositionNova 600Transfer 150Astaxanthin all-E (mg y kg )19.922.6Astaxanthin Z -isomers (mg y kg )7.47.6Cantaxanthin (mg y kg )34.238.6Semiastacene (mg y kg ) 1.40.9Lutein (mg y kg )7.79.0Zeaxanthin (mg y kg ) 6.78.1Protein (%)4245Fat (%)3530Energy (MJ y kg )25.524.3formance liquid chromatography (HPLC )of mus-cle extracts of Atlantic salmon fed diets with canthaxanthin.The second objective of the present work was to isolate and characterise this com-pound,and to quantify the amount present in the flesh of 0q and 1q smolts of Atlantic salmon fed a diet supplemented with 30mg y kg astaxan-thin and 30mg y kg canthaxanthin for 35weeks.Carotenoid composition,colour characteristics,growth and condition factor of the fish are also reported.2.Materials and methods2.1.Fish and experimental design of the feeding trialDuplicate netpens (25=25m )of Atlantic salm-on (S.salar ,NLA stock )0q smolts (stocked in seawater in November 2000)and 1q smolts (stocked in seawater in April 2001)were set up at Gratnes,Rogaland,Norway.All fish were taken ˚from the same generation and eggs were hatched in January 2000.An account of the selective breeding program of the NLA stock was recently given by Gjedrem (2000).After transfer to the sea,the fish were fed a diet supplemented with 30mg y kg astaxanthin and 30mg y kg canthaxanthin (first EWOS Transfer and then EWOS nova;EWOS Innovation,Dirdal,Nor-way )for 35weeks (from July 2001to February 2002).The diets were based on commercial feeds and the proximate compositions are shown in Table 1.Before transfer to the sea,salmon received a commercial diet without any supplemented carot-enoids.The fish were fed to satiation using an automatic feeding system with a Doppler sensor (Akvasmart,Bryne,Norway ).Every 8weeks,50fish per cage were caught at random and killed.The fish were organised into 5pools of 10per cage according to their live mass.The Norwegian Quality Cut (NQC )sample was taken from each fish according to NS 9401(Norwegian Standard,1994a )and the 10NQCs per pool were homogen-ised.A sub-sample of the homogenate from each pool was packed in a sealed plastic container and stored at y 208C.Analyses were performed within 4weeks after arrival at the laboratory to reduce potential loss of carotenoids.2.2.Reagents and chemicalsAll HPLC grade solvents for chromatography and solvents for extraction (pro analysis quality )were purchased from Merck (Darmstadt,Germa-ny ).Dry pyridine (H O -0.05%),triethylamine,2N -ethyl (diisopropyl )amine and (y )-camphanoyl chloride were purchased from Fluka Chemie (Buchs,Switzerland ).(S )-(q )-a -(1-Naphthyl )-ethyl isocyanate and ethylene glycol monoethyl ether were purchased from Sigma-Aldrich (Schnel-ldorf,Germany ).Analytical standards of crystal-line synthetic astaxanthin consisting of a 1:2:1mixture of the (3S ,39S )-,(3R ,39S ;meso )-,and (3R ,39R )-astaxanthin optical isomers and crystal-line synthetic canthaxanthin were obtained from Hoffmann-La Roche (Basel,Switzerland ).32T.Ytrestøyl et al./Comparative Biochemistry and Physiology Part B138(2004)29–402.3.Isolation and characterisation of49-hydro-xyechinenoneFillets from five commercially farmed Atlanticsalmon fed a diet supplemented with canthaxanthinand astaxanthin were supplied by Aquascot(Alness,Scotland,UK)and used as a source of 49-hydroxyechinenone.The fish were shipped fro-zen to AKV AFORSK,Sunndalsøra,Norway.Thefish were filleted and the carotenoids of approxi-mately5kg minced flesh was extracted withacetone:methanol(7:3)until the homogenate wascolourless.The solvent was evaporated and theremaining oil and water mixture was extractedwith diethylether after addition of saturated NaCl-solution.After evaporation of the diethyl ether,theremaining oil was saponified with KOH to removehydrolyzable lipids and astaxanthin.The carote-noids were extracted using diethyl ether,evaporat-ed,and dissolved in n-hexane.The carotenoidswere fractionated by open column chromatographyon silica gel(30=500mm,Kieselgel60,meshsize230–400,Merck,Darmstadt,Germany)usingincreasing amounts of acetone in n-hexane asmobile phase.The fraction containing49-hydrox-yechinenone was purified repeatedly by thin-layerchromatography(TLC)on silica gel plates(20=20cm,0.25mm thickness,DC-Alufolien, Kieselgel60,product no.7731,Merck,Darmstadt, Germany)using acetone y n-hexane(20:80)as a mobile phase.Before sample application the silica gel plates were purified by elution with metha-nol:chloroform:ethyl acetate(1:1:1).A UV–visi-ble spectrum was collected for the isolated fraction containing49-hydroxyechinenone(Shimadzu UV 260,UV–Visible recording spectrophotometer, Shimadzu,Japan)to determine absorption curvesand absorption maximum,l(in n-hexane),andmaxcarotenoid content.It was compared with authenticstandard of49-hydroxyechinenone on TLC asdescribed above and HPLC on a Spherisorb S5CN-4800nitrile column(Hichrom,Theale,Berkshire,UK;length250mm;internal diameter4.6mm;particle size5m m;mobile phase:n-hexane y iso-propyl acetate y acetone,88:8.5:3.5).Mass spectraof49-hydroxyechinenone were collected on a Fin-nigan MAT SSQ710instrument(Bremen,Germa-ny)equipped with a direct insertion solids probe(DIP)using positive electron impact(70eV,20–700amu y s,source temperature2708C).The DIP temperature program was308C for1min,100 8C y min until3008C,and3008C for8min.A standard of rhodoxanthin(49,59-didehydro-4,59-retro-b,b-carotene-3,39-dione)was used to opti-mize the DIP-MS before sample application.2.4.Preparation of standards and derivativesAn authentic standard of49-hydroxyechinenone was prepared from crystalline synthetic canthax-anthin by reduction with NaBH in dry diethyl4ether for10min.The reaction was monitored by TLC on silica gel plates as described above.The reaction was terminated by addition of water sat-urated with NaCl and the carotenoids were extract-ed with diethyl ether.The49-hydroxyechinenone fraction was isolated from isozeaxanthin(4,49-dihydroxy-b,b-carotene)and intact canthaxanthin by TLC on silica gel plates,and identified by UV–VIS spectroscopy.Authentic standards of the 39,49-cis and39,49-trans glycolic isomers of idox-anthin were prepared similarly by reduction of astaxanthin(Aas et al.,1997).To be able to separate the optical R y S isomers of49-hydroxyechinenone,diastereomeric carba-mates were prepared from an aliquot of the isolated sample(ca.0.2mg)after reaction with(S)-(q)-a-(l-naphtyl)ethyl isocyanate(2drops),in a mix-ture of chloroform(3drops)and triethylamine(2 drops),for24h.The samples were purified as described by Ruttiman et al.(1983),and analysed¨by HPLC after serial coupling of two silica gel columns(Spherisorb S5W-250A,length250mm, internal diameter 4.6mm,particle size5m m; Hichrom).The mobile phase was n-hexane y acetic acid isopropyl ester(88:12)with0.2%ethylene glycol monoethyl ether(v y v)and0.1%(v y v)N-ethyldiisopropylamine added,and flow rate 1.2 ml y min.Detection was at460nm.Retention times for the carbamates49R and49S -isomers were26.0and26.6min,respectively. However,the resolution was not satisfying and the R y S isomers were instead converted into diaster-eomeric camphanates prepared from(y)-cam-phanic acid chloride according to the method of Vecchi and Muller(1979).Optical isomer mixtures¨of49-hydroxyechinenone were converted into the corresponding esters of(y)-camphanic acid by reacting a dried sample of49-hydroxyechinenone with(y)-camphanoyl chloride in dry pyridine(1 ml)at0–48C.The relative amount of the two configurational isomers of49-hydroxyechinenone were determined quantitatively by HPLC on a Spherisorb S5CN-4800nitrile column according to33 T.Ytrestøyl et al./Comparative Biochemistry and Physiology Part B138(2004)29–40Vecchi and Muller(1979).Synthetic49-hydrox-¨yechinenone obtained from canthaxanthin consist-ing,of the(49R)-and(49S)-hydroxyechinenone-isomers in a1:1ratio,was used as a reference. The chromatographic resolution was improved,but the elution order was opposite to that of the carbamates.The retention times of the(49R)-and (49S)-isomers were14.3min and13.5min, respectively.2.5.Quantification of carotenoidsThe pooled homogenised samples were thawed, and carotenoids were extracted from samples of accurately weighed minced muscle using a1:1:3 mixture of distilled water,methanol(containing 500mg l BHT),and chloroform according to y1Bjerkeng et al.(1997a).The solvent was removed from an aliquot under reduced pressure,and re-dissolved in mobile phase(acetone y n-hexane y methanol20:80:0.1),filtered through a0.45m m filter(Minisart SRP15,Gottingen,Germany)and¨analysed isocratically by HPLC according to Bjer-keng et al.(1997a).The HPLC system used was a LC10AS pump connected to a SPD M6A photodiode array detector(detection wavelength at470nm)and a SIL10A autoinjector(Shimad-zu,Kyoto,Japan).Chromatogram re-integrations were performed using the LC Workstation Class-LC10software(Shimadzu,Kyoto,Japan).Con-centrations were calculated based on peak areas. Separations were achieved on a Spherisorb S5CN-4800nitrile column(Hichrom;length250mm; internal diameter4.6mm)using a flow of1.5ml y min,and a pressure of48–50bar.The retentiontimes(R)of canthaxanthin,49-hydroxyechineno-Tne,astaxanthin,and39,49-cis and39,49-trans gly-colic isomers of idoxanthin on this system were, 2.6,3.2,4.1,6.7and8.6min,respectively.All analyses were performed in duplicate.2.6.ColorimetryThe fish were treated as described above and filleted before instrumental tristimulus colour anal-ysis(CIE L*a*b*),CIE(1986).Muscle colour varies longitudinally in the fillets(No and Store-bakken,1991).Colorimetric measurements were, therefore,performed on the dorsal muscle posterior to the dorsal fin,under the adipose fin,and in the tail on each fillet,using a Minolta Chroma Meter CR-300(Minolta,Osaka,Japan),as described previously(Bjerkeng et al.,1997b).Measurements were made directly on the fillets,and the measur-ing head was rotated908between duplicate meas-urements per position,and means of six recordings per fish were used for data analysis.Visual colour assessments of the same fillets were also evaluated at the same positions using a Roche Salmo Fan(Hoffmann-La Roche,Basel, Switzerland).The colour was evaluated using the 20–34scale(20represents low degree of pigmen-tation and34represents a highly pigmented fillet). Colour card evaluations were carried out indoors by three trained panelists as described by Bjerkeng et al.(1997b).2.7.Determination of muscle lipidsThe content of crude lipids in the muscle was determined gravimetrically according to NS9402 (Norwegian Standard,1994b).Homogenized sam-ples of muscle(ca.5g)were dried with Na SO24 (10g)and extracted with ethyl acetate(50ml) by shaking in a water bath(258C)for1h.The extract was filtered,the solvent evaporated,and the lipid residue weighed.2.8.Data analysisSpecific growth rate(SGR)was calculated as100=(ln(W)y ln(W))y days of feeding,whereF IW and W indicate initial and final masses of theF Ifeeding period,respectively.Carotenoid accumu-lation efficiency was calculated as((C=W)-F F (C=W))y(W y W)where C and C indicateI I F I F Iinitial and final total carotenoid concentrations between consecutive weeks of sampling,respec-tively.Condition factor(CF)was calculated as body mass in g=(length in cm).Data werey3analysed statistically by analysis of covariance (ANCOV A)with smolt type(0q and1q)and sampling month as fixed factors and mass as covariate using the general linear models(GLM) statement of the SAS Software Release8.2for Windows(SAS Institute,Cary,NC,USA).One-way analyses of variance(ANOV A)were per-formed using smolt type or duration as factors. Significant(P-0.05)differences among means were ranked by Duncan tests and the results are presented as means"S.E.M.Percentages were arc-sin transformed before statistical analyses(Sokal and Rohlf,1995).Data on total carotenoid content were corrected for mass differences by using smolt34T.Ytrestøyl et al./Comparative Biochemistry and Physiology Part B 138(2004)29–40type (0q and 1q )as fixed factor and the natural logarithm of mass as covariate.Significant differ-ences between least square means were indicated when P -0.05.Regression models giving the best fit to the data were obtained using Microsoft Excel 9.0(Microsoft,Redmond,WA,USA ),and the non-linear regression models were tested using the NLIN procedure of the SAS Software.Linear relationships were tested using the REG procedure of the SAS Software.Significance was indicated when P -0.05.3.Results and discussion3.1.Characterisation of 49-hydroxyechinenone Suggested minimum criteria for identification of carotenoids require visible absorption spectra to determine l and possible spectral fine structure,max chromatographic properties in two different sys-tems (preferably TLC and HPLC )and co-chro-matography with authentic standard,and a mass spectrum to at least confirm molecular mass,all data to be consistent with the suggested structure (Schiedt and Liaaen-Jensen,1995).The VIS-spec-trum of the isolated 49-hydroxyechinenone revealed an echinenone-type chromophore typical for one keto-group conjugated to the chain of double bonds (l s 458nm in y n -hexane ).It co-chromatogra-max phied with authentic standard of 49-hydroxyechi-nenone obtained after NaBH reduction of 4canthaxanthin on both TLC (R s 0.3,20%acetone f in n -hexane )and HPLC (R s 3.2min ).The mass t spectrum exhibited a molecular ion at M s 566q (molecular mass;intensity 100%),and diagnostic fragment ions at 548(M -18;11%),474(M -q q 92;8%)and 460(M -106;3%).Small amounts q (-1%of total carotenoids )of a yellow xantho-phyll,presumably isozeaxanthin,were detected,but the structure of this carotenoid was not further elucidated.To determine the optical R y S isomer composi-tion of the 49-hydroxyechinenone isolated from Atlantic salmon,a purified sample was converted into the corresponding carbamates and analysedby HPLC according to Ruttiman et al.(1983).The¨chromatogram revealed a pronounced selective accumulation of the (49S )-hydroxyechinenone opti-cal isomer,cf.Fig.1.The (49S )-and (49R )-isomers comprised ca.81and 19%of the total 49-hydroxy-echinenone,respectively,as determined by thecamphanate method (Vecchi and Muller 1979).¨Thus,the selectivity of reduction of canthaxanthinto 49-hydroxyechinenone in Atlantic salmon appears to be less prominent than that reported for hens (Schiedt,1989),and that for skin of rainbow trout in which the (49S )-and (49R )-isomers com-prised ca.95and 5%of total 49-hydroxyechine-none,respectively (Bjerkeng et al.,1992).Thus enzymes,either isozymes or other enzymes with canthaxanthin reducing activity,capable of trans-forming canthaxanthin to 49-hydroxyechinenone may have different selectivity in different species and tissues.However,a selective metabolism or transport of (49R )-hydroxyechinenone may cause selective retention of (49S )-hydroxyechinenone in the muscle.Idoxanthin is an analogous reduction product of astaxanthin in Atlantic salmon (Schiedt et al.,1988a,b,1989)and Arctic charr (Aas et al.,1997;Hatlen et al.,1997),and has the same absolute configuration (49R ;Schiedt et al.,1988a )as the major part of the isolated 49-hydroxyechinenone.The regioselectivity of the reduction of the 4(49)-keto-groups of astaxanthin to idoxanthin was con-siderably higher than observed for canthaxanthin in the present experiment (Schiedt et al.,1988a ).Thus,introduction of 3(39)-hydroxy-gronps appar-ently increases the regioselectivity of the reaction and this may be explained by the steric require-ments of the enzyme.Nevertheless,the favoured reduction products of carotenoid substrates with 3-hydroxy-4-keto end-groups are the sterically hin-dered cis -glycolic forms (Bjerkeng et al.,2000).3.2.Growth,condition factor,carotenoid and lipid contents,and colour characteristicsManipulation of photoperiod is common prac-tice to alter timing of seawater transfer of Atlantic salmon smolts and facilitates production of mar-ketable fish with consistent size year round (Dun-can et al.,1998,2002).Groups transferred to seawater at different times of the year exhibit different growth profiles (cf.Duncan et al.,2002),and salmon transferred to seawater as 0q smolts have a higher growth rate during their life span than smolts transferred as 1q smolts.Similar growth profiles were observed in the present exper-iment (Fig.2),and the 0q smolts had significantly higher masses than the 1q smolts (P -0.001).The SGR dropped significantly from July–August to December–February for both 0q and 1q smolts (P -0.05),Fig. 3.Regression analysis35T.Ytrestøyl et al./Comparative Biochemistry and Physiology Part B 138(2004)29–40Fig.2.Mass (g )development of the 0q and 1q smolts of Atlantic salmon (Salmo salar )during the 35weeks feeding trial.Values are means "S.E.M.,n s100.Fig.3.Specific growth rates (SGR,%y day )for the 0q and 1q smolts of Atlantic salmon (Salmo salar )during the 35weeks feeding trial,n s100.Fig.4.Condition factor (g y cm )for the 0q and 1q smolts 3of Atlantic salmon (Salmo salar )during the 35weeks feeding trial,n s 100.showed a significant negative correlation between SGR and mass for both 0q (P -0.01;R s 0.75)2and 1q smolts (P -0.001;R s 0.64).However,2an ANCOV A using SGR as dependent variable,smolt type as fixed factor and mass as covariate did not show any significant differences in growth between 0q and 1q smolts (P s 0.41).The con-dition factor (CF )was significantly correlated to total body mass (P -0.01;R s 0.77and 0.82for 20q and 1q smolts,respectively ).An ANCOV A with CF as dependent variable,smolt type as fixed factor and mass as a covariate to correct for mass differences revealed that 0q smolt had signifi-cantly higher CF than 1q smolts (P -0.01),cf.Fig.4.A similar development in CF was observed by Mørkøre and Rørvik (2001).Our results indi-cate that timing of smolt production may have significant effects on body shape.Softer muscle texture has been associated with periods of high growth (Einen et al.,1999;Mørkøre and Rørvik,2001)and a firm muscle texture is correlated with a higher fibre density in the muscle (Johnston et al.,2000).Astaxanthin is deposited in the unesterified form in the muscle (Henmi et al.,1987),and the geometrical all-E isomer predominates ()90%of total astaxanthin;Henmi et al.,1990a;Bjerkeng et al.,1997a ).Canthaxanthin and astaxanthin are believed to be bound non-specifically to the hydrophobic pocket in the actomyosin protein-complex by weak inter-actions (Henmi et al.,1987,1989,1990b,1991).Since carotenoids are closely associated with mus-cle proteins,muscle growth rate or growth profilesmay influence pigmentation.There is,however,little information about the effects of growth pro-files on recruitment and growth of muscle cells,and the potential effects on deposition of carote-noids in the muscle.Johnston et al.(2000)com-pared an early and a late sexually maturing strain of Atlantic salmon,and found that muscle fibre density did not correlate significantly with astax-anthin concentration,whereas the visual colour scores correlated significantly with fibre density and were significantly different between strains.Thus,muscle structure may have a significant influence on appearance.In the present experi-ment,the 0q smolts had a higher total carotenoid content than the 1q smolts at all samplings (P -0.0001),Fig.5.The carotenoid accumulation pro-files indicated that the 0q smolt contained more carotenoids than the 1q smolt,also at a similar size.Fig.6.An ANCOV A with total carotenoids as dependent variable,smolt type as fixed factor and the natural logarithm of mass as a covariate to correct for mass differences revealed that 0q。