Growth, feed utilization and liver histology of juvenile common sole (Solea solea L.) fed isoenergetic diets with increasing protein levels

Growth, feed utilization and liver histology of juvenile

common sole (Solea solea L.) fed isoenergetic diets

with increasing protein levels

Pier Paolo Gatta1, Luca Parma1, Ilaria Guarniero2, Luciana Mandrioli2, Rubina Sirri2,Ramon Fontanillas3 & Alessio Bonaldo1

1Dipartimento di Morfo¢siologiaVeterinaria e Produzioni Animali, Bologna, Italy2Dipartimento di SanitaØ PubblicaVeterinaria e Patologia Animale, Bologna, Italy3Skretting Aquaculture Research Centre, Stavanger, Norway

Correspondence: A Bonaldo, Dipartimento di Morfo¢siologiaVeterinaria e Produzioni Animali,Via Tolara di Sopra 50, 40064 Bologna.

E-mail: [email protected]

Abstract

This study was undertaken to determine the in£u-ence of dietary protein levels on growth, feed utiliza-tion and liver histology in common sole. Fourisoenergetic diets were formulated to contain fourdi¡erent crude protein levels: 39 (P39), 45 (P45), 51(P51) and 57 (P57) % dry weight. Fifty animalsper tank (initial weight 10.2 � 0.4 g) were randomlydistributed into twelve 500 litre square tanks (bottomsurface: 5600 cm2) connected to a closed recircula-tion system. The diets were tested in triplicate for84 days. At the end of the experiment, the ¢nalweight ranged from 19.6 (P39) to 25.4 g (P57). Thespeci¢c growth rate showed statistical di¡erencesbetween groups, with the best results in the groupfed diet P57 (1.07% day�1). Signi¢cant di¡erencesbetween groups were also recorded for the feedconversion ratio, with values of 1.31, 1.28, 1.12 and0.94 in P39, P45, P51and P57 respectively. Gross lipide⁄ciency was also signi¢cantly a¡ected by thedietary treatment, with the highest value (42.07%)found in P57. Ammonia excretion, expressed asg100 g�1 protein intake, was signi¢cantly lowerfor group P39 (2.46) than groups P51 (4.70) andP57 (4.75). Increased incidence of lipid dropletsin hepatocytes was observed when the dietaryprotein levels increased and/or dietary lipiddecreased.

Keywords: common sole, growth, protein utiliza-tion, lipid e⁄ciency, liver histology

Introduction

Salmonids, gilthead sea bream (Sparus aurata L.) andEuropean sea bass (Dicentrarchus labrax L.) playa leadrole in European ¢sh production; however, otherfarmed ¢sh species are required to di¡erentiate andwiden the market supply. From emerging candidatespecies, Senegalese sole (Solea senegalensis, Kaup1858) and common sole (Solea solea L.) show promise,as stated formerly by Howell (1997) and more re-cently by Imsland, Foss, Conceic� a� o, Dinis DelbareSchram Kamstra Rema and White (2003). Hatcheryproduction of Solea spp has been accomplished quiteeasily, and this has been a common bottleneck in thecommercial productionof other potential marine ¢shspecies. However, it is during the juvenile stages thatseveral factors combine to reduce the growth perfor-mance and thereby reduce the potential for commer-cial farming activities. These factors include feedingbehaviour, susceptibility to disease and stocking den-sity (Day, Howell & Jones 1997; Imsland et al. 2003;Schram, Van der Heul, Kamstra & Verdegem 2006;Piccolo, Marono, Bovera, Tudisco Caricato & Nizza2008; SaŁ nchez, Ambrosio & Flos 2010). Focusing onthe nutrition and feeding of common sole juvenilesand ongrowing, which seem to be particularly criti-cal points, some signi¢cant steps forward have re-cently been achieved for S. senegalensis (Rema,Conceic� a� o, Evers, Castro-Cunha Dinis & Dias 2008;Borges, Oliveira, Casal, Dias Conceic� a� o & Valente2009; Rubio, Boluda Navarro, Madrid & SaŁ nchez-VaŁ z-quez 2009), establishing optimal protein and lipid

Aquaculture Research, 2011, 42, 313^321 doi:10.1111/j.1365-2109.2010.02622.x

r 2010 TheAuthorsAquaculture Researchr 2010 Blackwell Publishing Ltd 313

mailto:[email protected]

levels for growth. Other recent ¢ndings on juvenilesof S. senegalensis (Silva, Espe, Conceic� a� o, Dias & Va-lente 2009) and Solea aegyptiaca (Chabanaud 1927)(Bonaldo, Roem, Pecchini, Grilli & Gatta 2006) de-monstrated the ability of Solea spp. to grow equallywell when fed diets containing vegetable proteins inthe partial substitution of ¢shmeal. Less informationis currently available for common sole (S. solea) nutri-tion (Piccolo et al. 2008) and it is inappropriate to relyon data from Senegalese sole, given di¡erences in thegrowth performance, optimal thermal regime,broodstock behaviour and natural range (Imslandet al. 2003; Palazzi, Richard, Bozzato & Zanella2006). Because dietary energy and protein are recog-nized as key factors in£uencing both adequate ¢shnutrition and feeding costs (Watanabe 2002), andconsidering the lack of speci¢c knowledge on S. solea,the ¢rst aim of this researchwas to assess growth re-sponse and feed utilization on feeding common solejuveniles isoenergetic diets with di¡erent protein le-vels. Furthermore, considering the importance ofthe liver as an indicator of the nutritional and physio-logical status of ¢sh (Bell, Tocher, MacDonald & Sar-gent 1995; Robaina, Moyano, Izquierdo, Socorro,Vergara & Montero1997; Caballero, Lo¤ pez-Calero, So-corro, Roo, Izquierdo & Fe¤ rnandez 1999), the secondpurpose of this trial was to assess common sole liverhistology in response to the experimental diets.

Materials and methods

Diets

Four isoenergetic diets (estimated gross energy:23.5MJ kg�1 dry matter (DM)] were formulated tocontain di¡erent protein concentrations: 39 (P39),45 (P45),51 (P51) and 57 (P57) % DM.The energy le-vel was chosen based on those used in previous trialson juvenile Senegalese sole (Dias, Rueda-Jasso, Pan-serat, Conceic� a� o, Gomes & Dinis 2004; Rema et al.2008). Protein sources of the diets were representedby ¢shmeal and vegetable protein ingredients suchassoybean protein concentrate and wheat gluten mealat an increasing ratio.Their increase was in the sameproportion at each step in order to maintain the ami-no acid ratio constant among the diets. In the ab-sence of speci¢c data on the vitamin, mineral andtrace mineral requirements of common sole, require-ment data for other species were applied (NationalResearch Council 1993). The ingredients and theproximate composition are given inTable1. Pellets of4mm were produced using an experimental extru-

der by the Skretting Aquaculture Research Center,Stavanger, Norway. Pellets were crumbled and sievedto obtain suitable particle sizes.

Fish, experimental set-up and sampling

The experiment was carried out at the Laboratory ofAquaculture, Facultyof VeterinaryMedicine, Cesena-tico, Italy. The common sole (S. solea) juveniles withan initial average weight 10.2 � 0.4 g were obtainedfrom the hatchery Solea B.V, IJmuiden, the Nether-lands. Before the experiment, the ¢sh were acclima-tized for 4 weeks to the experimental facilities andfed commercial ¢shmeal-based diets (Skretting,Vero-na, Italy; crude protein 55%, crude fat 18%). At thestart of the trial, 50 ¢sh were randomly distributedinto each of twelve 500 litre square tanks (bottomsurface: 5600 cm2). Each diet was fed to triplicatetanks for 84 days. Tanks were provided with naturalseawater and connected to a unique closed recircula-tion system consisting of a mechanical sand ¢lter, anultraviolet light and a bio¢lter. The water exchangerate per tank was 100% every 2 h. The overall waterrenewal of the systemwas 5% daily.Temperaturewasmaintained constant at 20 � 1 1C throughout the ex-periment; the photoperiod was held constant at a12-h day length through arti¢cial light (200 lx atthe water surface ^ Delta Ohm luxmeter HD-9221,Delta-Ohm, Padua, Italy).Water temperature and dis-solved oxygen (� 6.5mg L�1) were monitored dailyin each tank. Ammonia (total ammonia nitrogen� 0.1mg L�1), nitrite (NO2

� � 0.2mg L�1) and ni-

Table 1 Ingredients and proximate composition of the ex-perimental diets

Diet

P39 P45 P51 P57

Dietary ingredients (%)

Fishmeal LT 26.7 32.0 34.7 40.0

Soybean protein concentrate 10.7 12.8 13.9 16.0

Wheat 31.1 24.5 21.0 14.4

Fish oil 19.0 16.2 14.7 11.9

Wheat gluten meal 10.7 12.8 13.9 15.9

Vitamin and mineral premix� 1.8 1.8 1.8 1.8

Proximate analysis

Dry matter (DM) (%) 91.8 93.2 92.7 91.4

Crude protein (% DM) 39.0 45.2 51.0 56.6

Crude fat (% DM) 23.3 22.8 20.4 18.1

Ash (% DM) 6.1 6.4 7.0 7.6

Gross energy (MJ kg�1) 23.5 23.5 23.8 23.0

Crude protein/gross energy (kg MJ�1) 16.6 19.2 21.4 24.6

�Skretting standard vitamin and mineral premix.

Dietary protein levels and Solea solea performance P P Gatta et al. Aquaculture Research, 2011, 42, 313^321

r 2010 TheAuthors314 Aquaculture Researchr 2010 Blackwell Publishing Ltd, Aquaculture Research, 42, 313^321

trate (NO3� � 50mg L�1) were determined spectro-

photometrically once weekly (Spectroquant Nova 60,Merk, Lab business) at 12.00 p.m. At the same time,pH (7.8^8.2) and salinity (28^33 g L�1) were also de-termined. Fishwere hand-fed twice daily (at 9.00 a.m.and 5.00 p.m.) at a ¢xed rate of 1.3% bodyweight day�1, 7 days week�1. According to the ap-petite of ¢sh registered in previous trials conductedin our facilities (unpubl. data), this ratio was consid-ered to be close to satiation. The feeding levels wererecalculated daily for each tank according to the feedconversion ratio (FCR) and speci¢c growth rate (SGR)obtained at each intermediate weighing. Feed losseswere minimal throughout the trial but, when neces-sary, the remaining feed was estimated and deductedfrom the feed intake for the overall calculations. Atthe beginning and at the end of the experiment, allthe ¢sh of each tank were individually weighed andthe total length was recorded. The total biomass wasalso determined at day 28 and 56 by bulk weighing.Carcass proximate composition was determined atthe beginning and at the end of the trial. In the for-mer case, ¢ve pooled samples of ¢ve ¢sh each weresampled to determine the initial proximate composi-tion, while in the latter case, one pooled sample of ¢ve¢sh from each tank was collected to determine the ¢-nal proximate composition. Furthermore, at the endof the trial, wet weight, viscera and liver weight wereindividually recorded from ¢ve ¢sh per tank for thedetermination of visceral and hepatosomatic indices.At this time, two ¢sh per tank were also randomlysampled for liver histology.In order to evaluate nitrogen metabolism, after the

end of the growth trial, ammonia excretionwas mea-sured in all tanks during a 24-h cycle, integrating re-peated collected values according to the followingformula given by Kaushik (1980):

Et ¼ V0DCþ CtDW

whereV0 is the volume of water in the tank, DC thedi¡erence in total nitrogen ammonia (Ci�Ci� t), Ct

the mean of the total nitrogen ammonia concentra-tion between two consecutive intervals (Ci1Ci� t/2),DW the £ow rate/unit of time ‘t’, t the unit of incre-ment in time inwhich concentrationvariation is con-sidered to be minimal and Et the total nitrogenammonia excreted by ¢sh per unit of time retainedThe water in£ow was held constant at 250 L h�1

in each tank the day before and during ammoniasampling.Water of each tank was sampled from theoutlet at 0, 2, 4, 6, 8, 10,12,18 and 24 h after the ¢rstmeal. A tank without ¢shwas used as a blank.

Samples were immediately stored at �32 1C untilanalysis. The ammonia concentration in the sampleswas measured using the indophenol method (Koro-le¡ 1983), and the overall data were expressed as g oftotal nitrogen ammonia per100 g protein intake.All experimental procedures were evaluated and

approved by the Ethical-scienti¢c Committee forAni-mal Experimentation of the University of Bologna, inaccordance with the European Community Councildirective (86/609/ECC).

Analyses of diets and body composition

The experimental diets and carcasses were analysedfor DM (drying to a constant weight in a stove at105 1C), crude protein (N � 6.25, determined usingthe Kjeldahl method), fat (Folch, Lees & Sloane Stan-ley1957) and ash content (incineration to a constantweight in a mu¥e oven at 450 1C).

Calculations

The formulae used were calculated as follows:

SGR (%day�1)5100 � (ln FBW� ln IBW)/days,where FBWand IBW represent the ¢nal and the initialweights (tank means) respectively. FCR5 feed given/weight gain. Condition factor (CF)5100 � (bodyweight/total length3). Viscerosomatic index (VSI)(%)5100 � (viscera weight/body weight). Hepatoso-matic index (HSI) (%)5100 � (liver weight/bodyweight). Protein e⁄ciency ratio (PER)5body weightgain/protein intake. Gross protein e⁄ciency (GPE) (%)5100 � [(% ¢nal body protein � ¢nal body weight)� (% initial body protein � initial body weight)]/totalprotein intake ¢sh. Gross lipid e⁄ciency (GLE)(%)5100 � [(% ¢nal body lipid � ¢nal body weight)� (% initial body lipid � initial body weight)]/totallipid intake ¢sh.

Liver histology

At the end of the trial, two ¢sh per tank weresampled for liver histology. Samples were ¢xed in10% bu¡ered formalin, dehydrated in a gradedethanol series and embedded in para⁄n. Sectionseries of 4 mm were stained with haematoxylinand eosin (H&E). From each ¢sh, a sample of liverwas snap-frozen in liquid nitrogen. Quenching wasaccomplished by placing a small amount of embed-ding medium for frozen tissue specimens (OCT)

Aquaculture Research, 2011, 42, 313^321 Dietary protein levels and Solea solea performance P P Gatta et al.

r 2010 TheAuthorsAquaculture Researchr 2010 Blackwell Publishing Ltd, Aquaculture Research, 42, 313^321 315

(Tissue-Tek, Sakura ¢netek, Torrance, CA, USA) ontoa cork disc where the sample was positioned, andthen dropping it into a beaker containing isopentaneand liquid nitrogen for 1min and stored at �80 1C.Frozen sections of 3 mm were cut in cryostate (LeicaMicrosystems GmbH,Wetzlar, Germany) and stainedwith Oil red O. Histological samples were evaluatedobjectively by two pathologists (R.S., L.M.) andblindly with respect to the diet group; liver sectionswere scanned at � 40 lens with a light microscopeand10 ¢elds were selected; according to the grade ofseverity and the distribution of the fatty in¢ltrationwithin the cells (hepatic steatosis), and through thehistological section (from multifocal to di¡use distri-bution), cases were classi¢ed as mildly a¡ected (cyto-plasmic ¢lling of a clear, optically empty contentforming microvesicular spaces), moderately a¡ected(cytoplasmic ¢lling of a clear, optically empty contentpulling the nucleus at the cell periphery and formingmacrovesicular spaces) or severelya¡ected (cytoplas-mic ¢lling of a clear, optically empty content, confer-ring an appearance of signet-ring cells). If thesechanges were not present, the livers were consideredto be normal.

Statistical analyses

The performance response and the proximate com-position of ¢sh fed the experimental diets were ana-lysed using nested ANOVA and the Newman^Keulspost hoc test. The presence of morphological changes

in the distal intestinal structure of di¡erent groups of¢sh at the end of the trials was compared using theKendall t rank correlation coe⁄cient. All statisticalanalyses were performed using SAS computer soft-ware (SAS 2004). Tank was used as the experimentalunit for analyzing the growth performance and am-monia excretion; a pool of ¢ve sampled ¢sh was con-sidered to be the experimental unit for analysingcarcass composition, whereas individual ¢sh wasused for analysing CF,VSI andHSIandmorphologicalchanges. Signi¢cant di¡erences were assumed whenP � 0.05.

Results

All ¢sh readily accepted the experimental diets andfeed intakewas determined directly by the feeding re-gime. The overall mean mortality in terms of thenumber of ¢sh was o2% and no signi¢cant di¡er-ences were recorded between groups (P 50.9707).Growth and feed utilization are shown in Table 2.The ¢nal weight, weight gain and SGR in ¢sh fed dietP57were signi¢cantlyhigher than those found in theother groups. Similarly, ¢sh fed diet P57 had signi¢-cantly lower FCR in comparison with the other ¢shgroups. Fish fed diets P39 and P45 showed similarresults, while the group fed diet P51had intermediatevalues, with statistical di¡erences for weight gain,SGR and FCR.Data on nutrient retention e⁄ciencyand ammonia

excretion are presented in Table 2. Protein e⁄ciency

Table 2 Growth, feed utilization and ammonia excretion of sole fed with experimental diets

Parameters

Diet

P39 P45 P51 P57

IBW 10.1 � 0.3 10.1 � 0.7 10.2 � 0.3 10.3 � 0.3

FBW 19.6 � 0.2a 20.0 � 1.6a 21.6 � 1.2a 25.4 � 0.3b

WG 9.6 � 0.4a 9.9 � 0.9a 11.4 � 0.9b 15.1 � 0.3c

SGR 0.80 � 0.04a 0.81 � 0.02a 0.89 � 0.04b 1.07 � 0.03c

FCR 1.31 � 0.02a 1.28 � 0.01a 1.12 � 0.07b 0.94 � 0.02c

PER 1.93 � 0.10 1.73 � 0.03 1.73 � 0.09 1.86 � 0.03

GPE 35.06 � 0.42 29.68 � 5.25 30.33 � 1.80 33.72 � 7.82

GLE 26.23 � 2.50a 27.29 � 4.60a 32.19 � 3.37ab 42.07 � 5.25b

Total ammonia nitrogen excretion 2.46 � 0.26a 3.51 � 0.21ab 4.70 � 0.89b 4.75 � 0.23b

Data (mean � SD, n 53) in the same row with di¡erent superscript letters are signi¢cantly di¡erent (P � 0.05).IBW, initial body weight (g); FBW, ¢nal body weight (g); WG, weight gain (g) 5 FBW� IBW; SGR, speci¢c growth rate (% day�1)5100 � (ln FBW� ln IBW) days�1; FCR, feed conversion rate 5 feed given/WG; PER, protein e⁄ciency ratio5body weight gain/pro-tein intake; GPE, gross protein e⁄ciency (%) 5100 � [(% ¢nal body protein � ¢nal body weight)� (% initial body protein � initialbody weight)]/total protein intake ¢sh; GLE, gross lipid e⁄ciency (%) 5100 � [(% ¢nal body lipid � ¢nal body weight)� (% initialbody lipid � initial body weight)/total lipid intake ¢sh)]; total ammonia nitrogen excretion (g100 g�1 protein intake)5100 � [(totalammonia nitrogen (g)/protein intake (100 g)].



ratio and GPE were not statistically di¡erent amonggroups. Gross lipid e⁄ciency was higher in ¢sh feddiet P57 and signi¢cantly di¡erent from ¢sh fed dietsP39 and P45. Ammonia excretion was not statisti-cally di¡erent between groups fed diets P45, P51andP57 and signi¢cant di¡erences were only recordedbetween group P39 and both P51 and P57. Carcassproximate composition and biometric parameterswere very similar between groups, with no signi¢-cant di¡erences (Table 3). According to the histologyobservation, the livers were pink or white, and therich vascular network, which constitutes a physiolo-gical feature of these ¢sh, was evident. Twelve cases,which did not show intracytoplasmic lipidic dropletsin the hepatocytes, both on H&E and Oil red O stain-ing, were classi¢ed as normal (Fig.1a^d). In six cases,there was a cytoplasmic ¢lling of a clear, opticallyempty content forming microvesicular spaces (mildhepatic fatty in¢ltration); in these cases, Oil red Ostain demonstrated small, intracellular orange dro-plets (Fig.1e and f). In the other six cases, this materi-al, markedly pulling the nucleus at the cell periphery,showed a macrovesicular appearance to the hepato-cytes, and showed a multifocal to di¡use distributionthrough the histological section (moderate, multifo-cal to di¡use hepatic fatty in¢ltration) (Fig. 1g); Oilred O stain showed larger orange droplets. Occasion-ally, very large extracellular orange droplets werefound, due to the rupture of cytoplasmic membranesduring freezing and condensation of the lipidic con-tent (Fig.1h). In all cases, a mild, di¡use cytoplasmicswelling was detected (hepatic hydropic degenera-tion). All these ¢ndings were statistically more fre-quent in ¢sh fed diets P51and P57 than in those feddiets P39 and P45.

Discussion

While the growth of common sole has been studiedfor many years (Howell 1997) and many attemptshave been made to overcome low growth rates, poorresults have been achieved so far. Sole have two unu-sual features that may limit growth potential. Theyhave a peculiar gut morphology, which presents a re-latively small stomach without pyloric caecae, and along intestine; furthermore, feeding behaviour in thenatural environment occurs almost entirely bymeans of chemoreception and with a frequent inges-tionof small prey items (de Groot1971). In juveniles ofSenegalese sole fed isoenergetic diets with ¢ve di¡er-ent protein levels (CP 43; 48; 53; 57; and 60), Remaet al. (2008) found SGR ranging from 0.93 to1.22% day�1 with statistical di¡erences related todietary protein levels. In particular, at least 53% CPwas necessary to obtain the maximum growth rate.In the present trial, SGR ranged from 0.80 to1.07% day�1. Speci¢c growth rate was signi¢cantlya¡ected by the dietary protein level, with the highestvalues recorded for ¢sh fed diet D57. These ¢ndings,although the experiment was shorter, are compar-able to the growth rate described by Howell (1997) incommon sole, indicating that diet P57 is able to sus-tain good performance of S. solea juveniles.Concerning feed utilization, the present study de-

monstrated the in£uence of dietary protein levels onFCR and GLE but not on PER and GPE. A comparisonwith other data on common sole is quite di⁄cult be-cause of the paucity of sources. Piccolo et al. (2008)evaluated the performance of 30 g common sole feddiets containing 50% (diet A) and 54% (diet B) crudeprotein for 300 days. The FCRwere 2.65 and 2.49 for

Table 3 Whole body proximate composition and biometric parameters of sole fed with the experimental diets

P39 P45 P51 P57

Proximate composition

Moisture 75.35 � 0.38 75.20 � 0.58 75.56 � 0.19 75.23 � 0.88

Protein 16.56 � 0.35 16.05 � 1.58 16.39 � 0.78 16.79 � 2.40

Lipid 6.93 � 0.38 6.85 � 0.61 6.67 � 0.35 6.61 � 0.49

Ash 2.83 � 0.07 2.48 � 0.20 2.17 � 0.07 2.39 � 0.45

Biometric parameters

CF 1.04 � 0.12 1.05 � 0.13 1.04 � 0.13 1.07 � 0.12

VSI 4.88 � 0.48 5.04 � 0.59 5.31 � 0.95 5.42 � 0.77

HSI 1.07 � 0.22 1.06 � 0.18 1.11 � 0.26 1.19 � 0.17

Data are shown as mean � SD.Moisture, protein, lipid and ash (% wet weight), n 51pool of ¢ve ¢sh per tank; CF, condition factor 5100 � (body weight/total length3),n550 per tank; VSI, viscerosomatic index (%)5100 � (viscera weight/body weight), n510 per tank; HSI, hepatosomatic index(%) 5100 � (liver weight/body weight), n 510 per tank.



diets A and B, respectively, with a statistical di¡er-ence between treatments. The authors reported thepresence of uneaten feed throughout the experimen-tal period and hence those values could not representthe real FCR. Rema et al. (2008) carried out a trial onSenegalese sole juveniles to de¢ne the optimal pro-tein levels with experimental diets containing acrude protein percentage of 43, 48, 53, 57 and 60.They found feed e⁄ciency very close to one for thelast three groups. In the present trial, FCR decreased

progressively from 1.31 in ¢sh fed the lowest proteinlevel to 0.94 in ¢sh fed the highest protein level. Inthe current trial, the highest protein level (diet P57)resulted in the lowest FCR. This was lower than thatreported previously for Senegalese sole fed optimaldietary protein levels (Rema et al. 2008). No data areavailable on protein utilization for common sole andhence a comparison canonly be madewith other ¢shspecies. The present study showed no in£uence ofdietary protein level on PER and GPE and this is inaccordance with previous ¢ndings obtained by Remaet al. (2008) for juvenile of Senegalese sole fed dietscontaining di¡erent protein levels ranging from 43%to 60%.Cadena-Roa (1983) studied the protein require-

ments of common sole using semi-puri¢ed dietsbased on casein and gelatin, with protein levels ran-ging from24% to 77%, and the best performancewasobtained at 57^58% dietary protein. This ¢ndingsupports the results of the present trial for ¢sh feddiet P57, where SGR and FCRwere not only the bestamong the other groups but also very similar to whatwas found for Senegalese sole. The data shown in thepresent study cannot clarify whether a higherdietary protein level could improve common solegrowth performance; however, it seems reasonable toconsider 57% crude protein to be a reference pointquite close to the requirements for juvenile commonsole. In addition to growth results, GLE increased sig-ni¢cantly with increasing dietary protein levels(Table 2). Because higher dietary protein levels werebalanced by a proportional lipid and carbohydratereduction (Table1) and assuming that higher proteinlevels led to good nitrogen utilization in this experi-ment, it is probable that lipid retention is mainlyrelated to the dietary lipid and carbohydrate levels,with a negligible in£uence of dietary protein levels.This hypothesis is in accordance with the resultsfound by Rema et al. (2008), where a lower dietarylipid content (between 10% and 13%) resulted inhigher lipid retention. These data may suggest that alower dietary lipid content than those used in thepresent trial is capable of ful¢lling common sole lipidrequirements, irrespective of the dietary proteinlevels, as already stated for Senegalese sole (Remaet al. 2008; Borges et al. 2009).Whole-body composi-tion (Table 3) was in accordance with the valuesfound for Senegalese sole (Dias et al. 2004; Remaet al. 2008). Surprisingly, the proximate carcass com-positionwas not a¡ected by the dietary treatments incontrast to Rema et al. (2008), who found that in-creased dietary protein levels tended to decrease

Figure 1 (a^d) Hepatic parenchyma from sole fed dietP39 (a and b) and diet P45 (c and d) was considered to be‘Normal’. Absence of intracytoplasmic lipidic droplets inthe hepatocytes, both on H&E (a, c) and Oil red O stainings(b, d), � 40 lens. (e and f) Hepatic parenchyma from solefed diet P51was considered to have mild hepatic fatty in-¢ltration. Hepatocytes contain a clear, optically emptycontent forming microvesicular intracytoplasmic spaces(e). Oil red O stain demonstrated small, single, intracellu-lar orange droplets (f), � 40 lens. (g and h) Hepatic par-enchyma from sole fed diet P57 was considered to havemoderate hepatic fatty in¢ltration. Hepatocytes are ¢lledwith clear, optically empty content that pulls the nucleusat the cell periphery and gives a macrovesicular appear-ance (g). Oil red O stain detected large orange droplets (h),� 40 lens.



whole-body fat deposition. The carcass proximatecomposition in the present study is similar to pre-viously reported data for other £at¢sh species, suchas Atlantic halibut (Aksnes, Hjertnes & Opstvedt1996). Hepatosomatic index observed in this study,ranging from1.06 to1.19, are within the values foundin a trial conducted by Dias et al. (2004) with juvenileSenegalese sole, whereas VSI were slightly higher inthe present study.Liver histology showed a higher density of intracy-

toplasmic lipidic droplets (steatosis) in ¢sh fed dietsP51 and P57 (Table 4). Steatosis can be associatedwith an excess of lipids in the diet, resulting in fataccumulation in the liver (Rueda-Jasso,Conceic� a� o,Dias, De Coen, Gomes, Rees, Soares, Dinis & Sorge-loos 2004; Myers & Mc Gavin 2007), or to a de¢-ciency in lipotropic factors that are able to preventor remove an excessive accumulation of fat in the li-ver (Mato, Mart|¤ nez-Chantar & Lu 2008). In thiscase, the inverse correlation between dietary lipid le-vels and the presence of lipid droplets in hepatocytesled us to lean towards a third mechanism.The liver isthe major site of de novo fatty acid synthesis in ¢sh(Lin, Romsos,Tack & Leveille 1977; Henderson & Sar-gent, 1981) as in mammals (Hillgartner, Salati &Goodridge 1995), and it is proposed that the elevatedincidence of intracytoplasmic lipidic droplets for thehigh-protein/low-lipid diet is due to increased hepa-tic lipogenesis. This is supported by the elevated ac-tivity of lipogenic enzymes, i.e. FAS, G6PDH, ME and12 ACC, which have been described previously for avariety of ¢sh species fed similar diets, includingchannel cat¢sh (Likimani & Wilson 1982), carp (Shi-meno, Kheyyali & Shikata 1995), European sea bass(Dias, Alvarez, Diez, Arzel, Corraze, Bautista &Kaushik 1998), Atlantic salmon (Arnesen, Krogdahl& Kristiansen 1993) or rainbow trout (Kolditz,Borthaire, Richard, Corraze, Panserat, Vachot, Le-fevre & Medale 2008).

Conclusion

In conclusion, the dietary protein levels used in thistrial demonstrated a considerable in£uence ongrowth, feed utilization and nitrogen excretion incommon sole juveniles, with the highest SGR andthe lowest FCR achieved with a diet containing 57%crude protein (1.07% day�1 and 0.94 respectively).Lipid retention increased when the dietary proteinincreased and dietary lipids decreased. Furtherresearch is needed to clarify both the protein and

protein/energy ratio requirements of common solejuveniles and the protein level in£uences on GLE.Regarding liver histology, the presence of lipiddroplets in hepatocytes suggested a lipogenesis en-hancement when the dietary protein levels increaseand/or the dietary lipids decrease. Finally, the overallperformances of common sole juvenile, especiallythose related to the growth and the FCR of groupP57, allowus to con¢rm the good potential of this ¢shspecies for intensive aquaculture.

Acknowledgments

This research was supported by grants from the Ita-lian Region Emilia-Romagna.We thank LorenzoMar-iani, Marina Silvi and Sara Giuliani for technicalassistance, Elettra Pignotti for statistical analysesand Leo Nankervis for English editing. The dietswere kindly provided by Skretting ARC, Stavanger,Norway.

References

Aksnes A., Hjertnes T. & Opstvedt J (1996) E¡ect of dietaryprotein level on growth and carcass composition inAtlantic halibut (Hippoglossus hippoglossus L). Aquaculture145, 225^233.

Arnesen P., Krogdahl —. & Kristiansen I.�. (1993) Lipogenicenzyme activities in liver of Atlantic salmon (Salmo salar,L). Comparative Biochemistry and Physiology B 105,541^546.

Bell J.G., Tocher D.R., MacDonald F.M. & Sargent J.R. (1995)E¡ects of dietary borage oil [enriched in g-linolenicacid,18: 3(n-6)] or marine ¢sh oil [enriched in eicosapen-taenoic acid,20:5(n-3)] on growth, mortalities, liver histo-pathology and lipid composition of juvenile turbot(Scophthalmusmaximus). Fish Physiology andBiochemistry14,373^383.

Table 4 Number of animals per dietary treatment thatdisplayed of morphological changes in liver (n56)

Diet Normal

Mild lipidicmorphologicalchanges

Moderate lipidicmorphologicalchanges

P39 6 0 0

P45 5 1 0

P51 0 3 3

P57 1 2 3

t 0.421

P 0.019



Bonaldo A., Roem A.J., Pecchini A., Grilli E. & Gatta P.P.(2006) In£uence of dietary soybean meal levels ongrowth, feed utilization and gut histology on Egyptiansole (Solea aegyptiaca) juveniles. Aquaculture 261,580^586.

Borges P., Oliveira B., Casal S., Dias J., Conceic� a� o L. & ValenteL.M.P. (2009) Dietary lipid level a¡ects growth perfor-mance and nutrient utilisation of Senegalese sole (Soleasenegalensis) juveniles. British Journal of Nutrition 102,1007^1014.

Caballero M.J., Lo¤ pez-Calero G., Socorro J., Roo F.J., Izquier-do M.S. & Fe¤ rnandez A.J. (1999) Combined e¡ect of lipidlevel and ¢shmeal quality on liver histology of giltheadseabream (Sparus aurata). Aquaculture179, 277^290.

Cadena-RoaM. (1983) Etude experimentale de l’alimentationde la sole (Solea vulgaris, Q.) en e¤ levage intensif,The' se Doc-torat de 3e¤ me Cycle Univ. de Bretagne Occidentale, Brest,France.

Day O.J., Howell B.R. & Jones D.A. (1997) The e¡ect ofdietary hydrolysed ¢sh protein concentrate on thesurvival and growth of juvenile Dover sole, Solea solea(L.), during and after weaning. Aquaculture Research 28,911^921.

Dias J., Alvarez M.J., Diez A., Arzel J., Corraze G., BautistaJ.M. & Kaushik S.J. (1998) Regulation of hepatic lipo-genesis by dietary protein/energy in juvenileEuropean seabass (Dicentrarchus labrax). Aquaculture161,169^186.

Dias J., Rueda-Jasso R., Panserat S., Conceic� a� o L.E.C., GomesE.F. & Dinis M.T. (2004) E¡ect of dietary carbohydrate-to-lipid ratios on growth, lipid deposition and metabolic he-patic enzymes in juvenile Senegalese sole (Solea senegalen-sis, Kaup). Aquaculture Research 35,1122^1130.

Folch J., Lees M. & Sloane Stanley G.G. (1957) A simple meth-od for the isolationand puri¢cationof total lipids froman-imal tissues. Journal Biological Chemistry 226, 497^509.

de Groot S.J. (1971) On the inter-relationships between themorphology of the alimentary tract, food and feeding be-havior in £at¢shes Pisces: Pleuronectiformes. NetherlandJournal of Sea Research 5,121^196.

Henderson R.J. & Sargent J.R. (1981) Lipid biosynthesis inrainbow trout, Salmo gairdnerii, fed diets of di¡ering lipidcontent. Comparative Biochemistry and Physiology C 69,31^37.

Hillgartner F.B., Salati L.M. & GoodridgeA.G. (1995) Physio-logical and molecular mechanisms involved in nutri-tional regulation of fatty acid synthesis. PhysiologicalReview 75, 47^76.

Howell B.R. (1997) A re-appraisal of the potential of the sole,Solea solea (L.), for commercial cultivation. Aquaculture155,355^365.

Imsland A.K., Foss A., Conceic� a� o L.E.C., Dinis M.T., DelbareD., Schram E., Kamstra A., Rema P. & White P. (2003) Areview of the culture potential of Solea solea andS. senegalensis. Reviews in Fish Biology and Fisheries 13,379^408.

Kaushik S.J. (1980) In£uence of nutritional status on the dai-ly patterns of nitrogen excretion in the carp (CyprinusCarpio L.) and the rainbow trout (Salmo gairdneri R.). Re-production Nutrition Development 20,1751^1765.

Kolditz C., Borthaire M., Richard N., Corraze G., Panserat S.,Vachot C., Lefevre F. & Medale F. (2008) Liver and musclemetabolic changes induced bydietary energycontent andgenetic selection in rainbow trout (Oncorhynchus mykiss).American Journal of Physiology ^ Regulatory, Integrativeand Comparative Physiology 294,1154^1164.

Korole¡ F. (1983) Determination of ammonia. In:Methods ofSeawaterAnalysis, 2nd edn. (eds. by R. Grassho¡, M. Ehr-hardt & R. Kromling), pp. 150^162. Verlag Chimic,Wein-heim, Germany.

Likimani T.A. & Wilson R.P. (1982) E¡ects of diet on lipo-genic enzyme activities in channel cat¢sh hepatic andadipose tissue adipose tissue. Journal of Nutrition 112,112^117.

Lin H., Romsos D.R.,Tack P.I. & Leveille G.A. (1977) In£uenceof diet on in vitro and in vivo rates of fatty acid synthesisin Coho salmon [Oncorhynchus Kisutch (Walbaum)]. Jour-nal of Nutrition107,1677^1682.

Mato J.M., Mart|¤ nez-Chantar M.L. & Lu S.C. (2008) Methio-nine metabolism and liver disease. Annual Review of Nu-trition 28, 273^293.

Myers R.K. & Mc Gavin M.D. (2007) Cellular and tissue re-sponses to injury. In: Pathologic Basis ofVeterinary Disease,4th edn. (eds. byM.D. Gavin & J.F. Zachary), pp.3^62. Mos-byYear Book, St. Louis, MO, USA.

National Research Council (NRC). (1993) Nutrient Require-ments of Fish. National Academy Press,Washington, DC,USA.

Palazzi R., Richard J., Bozzato G. & Zanella L. (2006)Larval and juvenile rearing of common sole (Solea soleaL.) in the Northern Adriatic (Italy). Aquaculture 255,495^506.

Piccolo G., Marono S., Bovera F., Tudisco R., Caricato G. &Nizza A. (2008) E¡ect of stocking density and protein/fatratio of the diet on the growth of Dover sole (Solea solea).Aquaculture Research 39,1697^1704.

Rema P., Conceic� a� o L.E.C., Evers F., Castro-Cunha M., DinisM.T. & Dias J. (2008) Optimal dietary protein levels in juve-nile Senegalese sole (Solea senegalensis). Aquaculture Nu-trition14, 263^269.

Robaina L., Moyano F.J., Izquierdo M.S., Socorro J.,VergaraJ.M. & Montero D. (1997) Corn gluten and meat and bonemeals as protein sources in diets for gilthead seabream(Sparus aurata): Nutritional and histological implications.Aquaculture157,347^359.

Rubio V.C., Boluda Navarro D., Madrid J.A. & SaŁ nchez-VaŁ z-quez F.J. (2009) Macronutrient self-selection in Solea sene-galensis fed macronutrient diets and challenged withdietary protein dilutions. Aquaculture 291,95^100.

Rueda-Jasso R., Conceic� a� o L.E.C., Dias J., De CoenW., GomesE., Rees J.F., Soares F., Dinis M.T. & Sorgeloos P. (2004) Ef-fect of dietary non-protein energy levels on condition and



oxidative status of Senegalese sole (Solea senegalensis) ju-veniles. Aquaculture 231, 417^433.

SAS. (2004) SAS Institute Inc, 2004. SAS Version 9.0. SAS In-stitute, Cary, NC, USA.

SaŁ nchez P., Ambrosio P.P. & Flos R. (2010) Stocking densityand sex in£uence individual growth of Senegalese sole(Solea senegalensis). Aquaculture 300, 93–101.

Schram E.,Van der Heul J.W., Kamstra A. & Verdegem M.C.J.(2006) Stocking density-dependent growth of Dover sole(Solea solea). Aquaculture 252,339^347.

Shimeno S., Kheyyali D. & Shikata T. (1995) Metabolic re-sponse to dietary lipid to protein ratios in common carp.Fisheries Science 61,977^980.

Silva J.M.G., Espe M., Conceic� a� o L.E.C., Dias J. & ValenteL.M.P. (2009) Senegalese sole juveniles (Solea senegalensisKaup,1858) grow equally well on diets devoid of ¢shmealprovided the dietary amino acids are balanced. Aquacul-ture 296,309^317.

Watanabe T. (2002) Strategies for further development ofaquatic feeds. Fisheries Science 68, 242^252.



Documents

Growth, feed utilization and liver histology of juvenile common sole (Solea solea L.) fed isoenergetic diets with increasing protein levels