J Appl Ichthyol 28 477-487

Page 1

Impact of dietary protein hydrolysates on skeleton quality and proteome

in Diplodus sargus larvae

By M. de Vareilles1,2*, N. Richard1*, P. J. Gavaia1, T. S. Silva1, O. Cordeiro1, I. Guerreiro1, M. Yufera3, I. Batista4,C. Pires4, P. Pousao-Ferreira5, P. M. Rodrigues1, I. Rønnestad2, K. E. Fladmark6 and L. E. C. Conceicao1

1CIMAR ⁄ CCMAR, Universidade do Algarve, Campus de Gambelas, Faro, Portugal; 2Department of Biology, University ofBergen, Bergen, Norway; 3Instituto de Ciencias Marinas de Andalucıa (ICMAN-CSIC), Cadiz, Spain; 4Instituto Nacional dosRecursos Biologicos (INRB ⁄ IPIMAR), Lisboa, Portugal; 5Instituto Nacional dos Recursos Biologicos (INRB ⁄ IPIMAR SUL),Olhao, Portugal; 6Department of Molecular Biology, University of Bergen, Bergen, Norway

Summary

In order to investigate the effects of dietary protein hydroly-

sates (PH) on larval growth performance, skeleton quality andproteome expression, triplicate groups of white seabream(Diplodus sargus) larvae were co-fed from first-feeding with live

feed and three microencapsulated diets differing in the molec-ular weight of their PH fraction (Control – inclusion ofCPSP-90; H – inclusion of a high amount in 0.5–30 kDa

hydrolysates; L – inclusion of a high amount in <0.5 kDahydrolysates). At 15 days after hatching (DAH), proteomeexpression changes were assessed in entire larvae by two-

dimensional gel electrophoresis and the quality of larvalskeleton was analysed at 28 DAH through double staining ofcartilage and bone. Dietary PH fractions tested affectedgrowth, the larvae fed diet L being significantly larger than

those fed diet H, but it did not affect the incidence of deformedlarvae, nor the number of deformities per fish. Two-dimen-sional analysis of larvae proteome allowed the detection and

the comparative quantification of a total of 709 protein spotshaving a pI between 4 and 7, around half of which had anexpression significantly affected by dietary treatment, the main

difference being between proteome of Control larvae withthose of both groups L and H. From these spots, 52 proteinsinvolved in diverse processes such as cytoskeletal dynamics,energetic, lipoprotein, amino acid (AA), and nucleotide

metabolisms, protein chaperoning and degradation, and signaltransduction, were identified. This study revealed that themolecular weight of the dietary protein hydrolysate fraction

had a minor impact on skeletal deformities in white seabreamlarvae, but affected growth performance and had a strongimpact on larvae whole body proteome.

Introduction

The production of live feed for marine fish larviculture iscostly, labour intensive and its replacement by the develop-ment of high quality artificial microdiets has been of majorfocus in recent years (Yufera et al., 1999, 2005; Murray et al.,

2010). Although early life stages have high potential growthrates (Conceicao et al., 1997), generally poor growth perfor-mance has been achieved in marine fish larviculture, unless live

feed is used. High growth rates require an abundant supply ofdietary amino acids (AA) for anabolic as well as energeticpurposes (Finn et al., 2002). However, the early life stages of

fish seem to have a limited capacity for digesting complex

proteins (Rønnestad et al., 2003). Modulating the solubilityand molecular size of proteins presented in artificial diets hasbeen shown to affect larval quality of different fish species

(Zambonino Infante et al., 2005). For instance survival,growth, development of the digestive tract and ⁄ or theoccurrence of skeletal deformities were improved by the

inclusion of dietary protein hydrolysates (PH) in larvae ofvarious fish species such as European seabass (ZamboninoInfante et al., 1997; Cahu et al., 1999), rainbow trout(Dabrowski et al., 2003), Atlantic cod and Atlantic halibut

(Kvale et al., 2009), and gilthead seabream (Gisbert, 2010).However, including PH in aquafeeds deserves further attentionas some inclusion levels of dietary peptides and ⁄ or free AA

(FAA) have been shown to have a detrimental effect on growthperformance (Cahu et al., 1999; Kolkovski and Tandler, 2000;Carvalho et al., 2004; Tonheim et al., 2005).

White seabream (Diplodus sargus) is a relatively new speciesin marine aquaculture production, partly due to its interestingmarket value and also because it seems that its availability in

the wild has been decreasing (Pousao-Ferreira et al., 2001;Santos et al., 2006). One of the main constrains for thecommercial production of this species, as indeed of othersparids, is the high incidence of skeletal malformations

(Saavedra et al., 2006). The high levels of osteological anom-alies that affect hatchery produced larvae are a major problemfor aquaculture, increasing costs and labour (Koumoundou-

ros, 2010). Recent studies indicate nutritional factors, anddietary protein in particular, as one of the causes involved inthe onset and development of these skeletal deformities

(Saavedra et al., 2010a), and it is reasonable to suppose thatintroducing dietary PH could improve larval quality, asobserved for other fish species (Tonheim et al., 2005; and

references therein; Savoie et al., 2011, and rerefences therein).Thus, in order to verify to what extent skeletal deformities as

well as growth performance of larval D. sargus can be affectedby the availability of dietary nitrogen, graded levels of fish PH

were introduced in the microdiets of the larvae and theskeleton and whole-body proteome analysed.

Materials and methods

Husbandry, experimental set-up and feeding protocol

White seabream (Diplodus sargus) eggs were obtained from the

Aquaculture Research Station of CRIPSul ⁄ IPIMAR (Olhao,Portugal). After yolk sac absorption, the larvae were randomly*These authors have contributed equally to this work.

J. Appl. Ichthyol. 28 (2012), 477–487� 2012 Blackwell Verlag, BerlinISSN 0175–8659

Received: September 23, 2011Accepted: March 08, 2012

doi: 10.1111/j.1439-0426.2012.01986.x

U.S. Copyright Clearance Centre Code Statement: 0175–8659/2012/2803–0477$15.00/0

Applied IchthyologyJournal of

Page 2

distributed into nine 200 L cylindro-conical fibreglass tanks(80 larvae L)1) so as to test three dietary treatments in

triplicate. This semi-closed re-circulating system consisted ofnatural sea water subjected to biophysical filtration and UV-irradiation prior to entering the tanks at an initial rate of

0.4 L min)1, slowly increased to a maximum of 1 L min)1

(30% h)1 water renewal) from 12 days after hatching (DAH).Minimal aeration was supplied through an air tube with outletat the bottom of the tank. Dissolved oxygen, water temper-

ature and salinity (6.7 ± 0.6 mg L)1, 19.2 ± 0.8�C, 35 ± 0.5,respectively) were closely monitored and adjusted as necessary,and all tanks were cleaned on a daily basis, the surface layer by

a skimmer. As long as live feed was given to the larvae, thewater was kept green with the microalgae Nannochloropsisoculata and Isochrysis galbana. The diurnal light : dark cycle

was set at 14:10 hours from controlled illumination providedby overhead fluorescent lamps. From 8 DAH, the photoperiodwas extended to 24 h of light until the end of the experiment.Larvae were fed with Brachionus plicatilis enriched with

Protein Selco (INVE Aquaculture, Belgium) from 3 to20 DAH. Artemia nauplii (AF, Salt Lake) from 12 to 20 andthen Artemia metanauplii (EG, Salt Lake) enriched with Easy

DHA Selco (INVE, Aquaculture, Belgium) from 21 to27 DAH. The amount of live prey supplied was graduallyreduced and replaced by inert diets in this period, being of little

dietary significance after 9 DAH. Live feed was supplied twicea day, 20–30 min after administrating the inert diets. Fromfirst-feeding, larvae were co-fed with one of three microencap-

sulated diets (Table 1) formulated to differ in the type of PHand tested in triplicates: diet Control (C), with a commerciallyavailable hydrolysed fish meal, CPSP-90 (composed approx-imately by 50% molecules <0.5 kDa and 50% between 0.5

and 30 kDa), at an inclusion level of 10%, as commonly usedin larval diets; diet High (H), containing 10% more largermolecular weight hydrolysates (0.5–30 kDa) than the other

diets; and diet Low (L), containing 10% more smallermolecular weight hydrolysates (<0.5 kDa, i.e., FAA, di- andtri-peptides) than diets C and H. These two hydrolysates

fractions were obtained from a commercial fish proteinhydrolysate kindly provided by COPALIS. They were

obtained by ultrafiltration in a Pellicon system using Biomaxmembranes (Millipore) with molecular weight cut-off of 30 and

0.5 kDa. The size of the dietary microparticles was <200 lmduring the first 15 days, then 200–400 lm. The diets were givenmanually four times a day until 8 DAH, and then eight times a

day to the end of the trial with the help of automatic dispensersat night.

Growth performance and skeletal deformities analysis

Samples of twenty larvae per tank were sampled at 2 and15 DAH for length and dry weight measurements. Survival

was assessed as the number of larvae alive at the end of thetrial, after withdrawing those removed for sampling.

Skeletal deformities were examined at 28 DAH, on 30 larvae

per tank, through double staining of cartilage and boneaccording to the method of Gavaia et al. (2000). Thus,cartilages were stained with Alcian Blue 8 GX (Sigma) andcalcified structures with Alizarin Red S (Sigma). Deformities

occurring in skeletal structures were identified, classified basedof the classification established by Favaloro and Mazzola(2006) for sharpsnout seabream (Diplodus puntazzo), and

quantified. Table 2 presents the alphanumeric codificationused to classify the skeletal deformities observed in the larvae.In order to facilitate the analysis of the data, the vertebral

column was divided into four areas: cephalic (1st to 6thvertebrae), pre-haemal (6th to 10th vertebrae), haemal (11th to21th vertebrae) and caudal (22nd to 24th vertebrae), and

codified by a letter. Numbers were used to codify the type ofabnormalities observed.

Analysis of larval proteins by two-dimensional electrophoresis (2-DE)

At 15 DAH, an equivalent amount of larvae from each tank ofa same dietary treatment was pooled to obtain 400 mg of

entire larvae per treatment. Samples were lysed by sonicationin 2 volumes of ice-cold buffer [7 MM urea, 2 MM thiourea, 4%(w ⁄ v) CHAPS, 0.3% (w ⁄ v) DTT, 0.6% (v ⁄ v) protease

inhibitor cocktail (Sigma-Aldrich)]. Homogenates were centri-fuged twice at 12 000 g for 10 min at 4�C to pellet insolublematerial. The resulting supernatants were depleted of non-protein contaminants using a ReadyPrep� 2-D Cleanup Kit

(Bio-Rad) and resuspended in ReadyPrep 2-D rehydra-Table 1Composition of the experimental diets (%, DM basis)

Diets

Control High Low

IngredientsAglonorse micro feed 28 21 21CPSP-90 10 0 0Cuttle fish meal 10 10 10Casein 11 6 6Na-alginate 8 8 8Bread yeast 6 6 6Dextrine 6 6 6Soy lecithin 6 6 6Fish oil 7 9 9Premix 4 4 4Vitamin C 3 3 3Vitamin E 1 1 1HYDROL_0,5 0 5 15HYDROL_0,5_30 0 15 5

Proximate composition (%)Dry matter 95.09 95.25 97.02Crude protein 47.62 46.65 46.72Crude fat 18.69 21.93 22.82Ash 6.24 6.80 6.53

Table 2Alphanumeric codification used to classify the skeletal deformitiesobserved in white seabream larvae at 28 DAH

Area Abnormalitya

A. Cephalic 3. Vertebral fusionB. Pre-haemal 4. Vertebral deformation (atrophy)C. Haemal 5. Malformed neural arch and ⁄ or spineD. Caudal 6. Malformed haemal arch and ⁄ or spineE. Pectoral fin 8. Malformed pterygophore (deformed,

absent, fused, supernumerary)F. Anal fin 9. Malformed parhypural ⁄ hypural (deformed,

absent, fused, supernumerary)G. Caudal fin 10. Malformed epural (deformed, absent,

fused, supernumerary)17. Malformed parapophyse18. Malformed left opercula19. Malformed right opercula20. Scolyosis

aAbnormalities 1, 2, 7, 11, 12, 13, 14, 15, 16 were removed since theywere not observed in white seabream larvae of this study

478 M. de Vareilles et al.

Page 3

tion ⁄ sample buffer (Bio-Rad). All protein quantificationswere performed using Quick Start� Bradford Protein Assay

(Bio-Rad).Larval proteins were first separated according to their

isoelectric point with ReadyStrip� IPG Strips, 11 cm long and

with a linear pH 4–7 (Bio-Rad). For each strip, 300 lg ofprotein in a final volume of 200 ll were loaded overnight bypassive rehydration. Isoelectric focusing was performed usingan Ettan IPGphor (GE Healthcare), at 20�C, for a total of

30 000 V.hrs. The remaining equilibration, second electropho-resis, staining and image analysis steps were performed as inAlves et al. (2010), using Criterion� XT Precast 10.5–14%

Tris-HCl gels (Bio-Rad) and an electrophoresis buffer con-taining 25 mMM Tris, 192 mMM glycine, 0.1% (w ⁄ v) SDS, pH 8.3.For each experimental condition, the 2-D gel electrophoresis

was performed in quadruplicates simultaneously.

Data analysis

Growth (dry weight, total length and growth rate), survivaland skeletal deformities data were subjected to one-way ANOVAANOVA

after confirming the normality and homoscedasticity assump-

tions and after arcsin transforming all data expressed aspercentages. When statistically significant variations werefound (P < 0.05), Tukey�s multiple comparison tests were

subsequently applied (P < 0.05). Growth rate was calculatedas 100�ðeðlnDWf�lnDWiÞ=t � 1Þ, with DWf and DWi as final andinitial dry weight (mg) of larvae, respectively, and t as number

of days. For length data, growth rate was obtained as100�ðTLf � TLiÞ=t , with TLf and TLi as final and initial totallarval length (mm), respectively, and t as number of days.A one-tailed non-parametric Wilcoxon-Mann–Whitney

U-test was used to assess differentially expressed proteinsbetween experimental groups. Protein spots that exhibited adifference in normalised volume superior to 1.5-fold between

experimental conditions, at P < 0.05, were considered to besignificantly differentially expressed. In addition, an explor-atory multivariate analysis was performed on this reduced set

containing only protein spots identified as differentiallyexpressed between experimental groups. All statistical analyseswere performed using the R software environment, version2.11.1 (R Development Core Team, 2011).

Protein identification by mass spectrometry

Protein spots differentially expressed between experimentalconditions that could be manually excised from preparativegels were identified after tryptic digestion by either LC-

MS ⁄ MS (ESI-Ion Trap) at the Aberdeen Proteomics facilities

(University of Aberdeen, UK) or MALDI Tof ⁄ ToF MS atthe Mass Spectrometry Laboratory, Analytical Services Unit

(ITQB, UNL, Portugal) and the Proteomics laboratory,Molecular Biology Department (University of Bergen, Nor-way). Details for the identification by LC-MS ⁄ MS have been

published elsewhere (Alves et al., 2010), as have those byMALDI ToF ⁄ ToFMS at the Proteomics laboratory (Gomez-Requeni et al., 2011). At the Mass Spectrometry Laboratory,the tryptic digests were dissolved in 2% acetonitrile ⁄ 5%formic acid and the extracted peptides loaded onto an R2microcolumn (RP-C18 equivalent) where they were desalted,concentrated and eluted directly onto a MALDI plate using a-ciano-hidroxycinamic (CHCA) as the matrix solution in 50%acetonitrile ⁄ 5% formic acid. Mass spectra of the peptideswere acquired with positive reflectron MS and MS ⁄ MS modes

using a 4800plus MALDI Tof ⁄ ToF� analyser (AB Sciex)with exclusion list of trypsin autolysis peaks (842.51, 1045.56,2211.11 and 2225.12). For both MALDI ToF ⁄ ToF MSresults, the collected MS andMS ⁄ MS spectra were analysed in

combined mode by using Mascot search engine and NCBI nr(ftp://ftp.ncbi.nih.gov/blast/db/README) database restrictedto Actinopterygii taxonomy, 50 ppm peptide mass tolerance

and 0.5 Da MS ⁄ MS mass tolerance, up to one missedcleavage allowed, formation of singly, positively chargedpeptides, carbamidoethylation of cysteine residues and possi-

ble oxidation of methionine residues.

Results

Growth and skeletal development response

There was a significant effect of diet on growth of whiteseabream larvae at 15 DAH but no differences were observedbetween survival rates by the end of the trial (Table 3). Fish

larvae fed diet H showed a higher growth rate, both in terms ofweight and total length, than those fed diet L, which showedthe lowest values among all groups.

The experimental trial was prolonged to 28 DAH for qualityassessment of larval skeleton, since the developmental state ofwhite seabream larvae at 15 DAH was at a stage too

premature to enable skeletal deformity analysis. Indeed, at15 DAH, most of the larvae only presented a few cartilaginousarches developing along the notochord and hypural elements

were not yet formed (data not shown). At 28 DAH, thedifferent PH fractions tested did not have any significant effecton the incidence of deformed white seabream larvae, whichwas 60 ± 19% in control group, 69 ± 13% in group H and

67 ± 12% in group L (mean ± standard deviation). Themaximum number of deformities recorded per larvae was 12but most of the deformed larvae presented <3 deformities.

Table 3Growth and survival of Diplodus sar-gus larvae fed a balanced diet with acommercially based fish protein hydro-lysate (Control), a balanced diet rich inlarger polypeptides: 0.5–30 kDa pep-tides – (High), and a balanced diet richin free amino acids, di- and tripeptides:<0.5 kDa – (Low). Values are mean± SEM (n = 20)

Control High Low

2 DAHDry weight (mg) 0.02 ± 0.01 0.02 ± 0.01 0.02 ± 0.01Total length (mm) 3.59 ± 0.03 3.59 ± 0.03 3.59 ± 0.03

15 DAHDry weight (mg) 0.06 ± 0.02ab 0.07 ± 0.02a 0.05 ± 0.00b

Total length (mm) 4.54 ± 0.09a 4.61 ± 0.08a 4.29 ± 0.09b

RGR (2–15 DAH)Dry weight (% day)1) 7.76 ± 0.91ab 9.42 ± 0.92a 7.01 ± 0.69b

Total length (% day)1) 7.37 ± 0.11a 7.84 ± 0.09a 5.45 ± 0.08b

Survival (%) 2.2 ± 0.3 2.7 ± 0.4 2.8 ± 0.2

Different letters in the same row indicate significant differences (One-way ANOVAANOVA and Tukey�s multiplecomparison tests, P < 0.05).

Dietary protein: skeleton and proteome in larvae 479

Page 4

The number of deformities per larvae was not statisticallysignificantly affected by dietary treatment (Fig. 1a). The

majority of the deformities were located in the cephalic areaof the vertebral column (Fig. 1b). The incidence of deformitiesin this region ranged from 51.4 to 72.7% of deformed larvae.

None of the vertebral column areas and fins analysed werefound to be significantly differentially affected by the incidenceof deformities depending on the dietary treatment. The mostfrequent type of abnormalities observed in all groups,

irrespective of the skeletal area affected, were malformedneural arches and ⁄ or spines, ranging from 66 to 81% ofdeformed larvae (Fig. 1c). A total of 20 different types of

deformities were recorded (Table 2). The most commondeformities found in all the groups were deformed neuralarches in cephalic (51.4–64.1% of deformed larvae) and pre-

haemal (17.2–43.3% of deformed larvae) vertebra (data notshown). No statistically significant differences between groupswere found concerning deformity types, though the incidenceof malformed parhypural ⁄ hypural in caudal fin tended to

develop less frequently in larvae fed diet H (1.4 ± 2.5% ofdeformed larvae) compared to those fed diet L (10.1 ± 5.5%of deformed larvae) (Tukey�s test, P = 0.065), and larvae of

group C tended to present less malformed neural arches in thehaemal area (4.5 ± 6.4% of deformed larvae) compared togroup H (19.6 ± 2.8% of deformed larvae) (Tukey�s test,

P = 0.059). Regarding severe types of deformities, scoliosis

was detected only in one individuals belonging to the controlgroup, and lordosis and kyphosis were never observed.

White seabream larval proteome response

The two-dimensional analysis of the white seabream larvaeproteome performed at 15 DAH for each dietary treatmentallowed the detection and the comparative quantification of atotal of 709 protein spots, of which 339 spots showed

significant variation depending on the diet (>1.5-fold or<0.67-fold, P < 0.05) (Fig. 2). Among these spots, 126 had asignificant differential expression between group C and both

groups L and H (100 spots under-expressed and 26 spots over-expressed in group C), 98 spots displayed significant variationsbetween groups C and L (45 spots under-expressed and 53

spots over-expressed in group L) and 84 between groups C andH (15 spots under-expressed and 69 spots over-expressed ingroup H). Proteome of larvae from groups H and L weresignificantly differentiated by the expression of 93 spots, of

which 44 were over-expressed in larvae fed the diet L and 49were under-expressed. Principal component (PC) analysis ofprotein spot expression data revealed a clear separation of

samples into clusters according to dietary treatment (Fig. 3),showing that the protein expression profiles of white seabreamlarvae fed with the experimental PH (diets H and L) were

rather similar, comparatively to that of larvae fed with the

0

10

20

30

40

50

60

70

0 1 2 3 >4

Def

orm

ities

inci

denc

e(%

tota

l lar

vae)

Number of deformities per larvae

C H L

0

20

40

60

80

100

G. Caudal finF. Anal finD. CaudalC. HaemalB. Pre-haemalA. Cephalic

Def

orm

ities

inci

denc

e (%

def

orm

ed la

rvae

)

Area affected

0

20

40

60

80

100

3 4 5 6 8 9 10 17 18 19 20

Def

orm

ities

inci

denc

e(%

def

orm

ed la

rvae

)

Abnormality code

(a)

(b)

(c) Fig. 1. Number of deformities perwhite seabream larvae incidence (%)(a), and distribution of deformed whiteseabream larvae according to (b) theskeletal area affected and (c) theabnormality type observed, at28 DAH depending on the dietarytreatment. Values are means with theirstandard deviation depicted by verticalbars (n = 30). Mean values were notsignificantly different between groups(P > 0.05). The control diet, diet Lowand diet High are labelled with C, Land H, respectively

480 M. de Vareilles et al.

Page 5

control diet. This was further evidenced by a heatmapgenerated from the relative abundance of identified proteinsfor all samples, in which protein spots were clearly groupedinto two main clusters, over- or under-expressed in group C

compared to remaining groups, and a fainter cluster within amain cluster, in which protein spots from group L were clearlyover-expressed compared to group C (Fig. 4).

Among the 52 altered proteins that could be excised andidentified (Table 4 and Fig. 4), 16 protein spots were charac-terised as cell cytoskeleton proteins, 9 of which were over-

expressed in group C (alpha and beta actins, tropomyosinalpha chains, myosin light chain 3 and type II keratins) and 7under-expressed in this group (alpha actins, cofilin 2, myosin-

binding protein H-like –MyBPH – and beta tubulin). The 13protein spots related to energy metabolic processes were eitherover-expressed in group L (fructose-bisphosphate aldolase –aldolase, glyceraldehyde-3-phosphate dehydrogenase –GAP-

DH, and enolases (-1 and -2), or in both groups H and L (ATPsynthase subunits, muscle-type creatine kinase isoformsand ⁄ or PTMs and enolase-3). Also over-expressed in groups

H and L were 2 AA metabolism proteins (glutamine synthetaseand antiquitin) and 2 proteins in nucleotide metabolism(nucleoside diphosphate kinase – NDK, and Apurinic ⁄apyrimidinic endonuclease – Ape, nuclear). The remaining 19protein spots identified were diverse functional proteins suchas crystallins (structure formation), proteasome subunits(proteasomal degradation pathway), pre-mRNA processing

factor 19 (RNA splicing factor, member of E3 ubiquitin ligasefamily), heat shock proteins (response to stress and chaperone

activity), 14-3-3 proteins (phosphoserine ⁄ phosphothreoninebinding, involved in many cellular pathways and stressresponse), protein disulfide isomerise – PDI (formation,reduction and isomerisation of disulphide bonds, chaperone

activity, lipoprotein metabolism and redox homeostasis), andcalbindin (calcium binding, many physiological processes).These were under-expressed in the control group compared to

groups H and L, except for one proteasome subunit (alphatype 3) over-expressed in group C, protein disulfide isome-rise –PDI over-expressed in group L relative to H, and

apolipoprotein A-IV4 (lipoprotein metabolism) over-expressedin group H.

(a)

(b)

Fig. 2. Two-dimensional gel electrophoresis of white seabream larvalwhole-body proteome showing protein spots which displayed statisti-cally significant variation between larvae fed the control diet andlarvae fed diets Low and High (a) and between larvae fed diet Low anddiet High (b). A molecular weight scale in kiloDalton (kDa) is shownto the left in (a). The pH range of the gel is of 4–7, from left to right. (a)Protein spots differentially expressed in control group compared to:both groups Low and High were marked with a cross; just group Low,with a square; and just group High, with a triangle. (b) Protein spotsdifferentially expressed in group Low compared to group H weremarked with a cross

–0.6 –0.4 –0.2 0.0 0.2 0.4 0.6

–0.4

–0.2

0.0

0.2

0.4

0.6

PC1

PC

2

C1C2

C3

C4L1

L2

L3

L4

H1H2

H3H4

–15 000 –5000 0 5000 10 000 15 000

–10

000

–500

00

5000

10 0

0015

000

SSP 0124

SSP 1201SSP 1327

SSP 1402

SSP 1405

SSP 1424

SSP 1704

SSP 2121

SSP 2309

SSP 2517

SSP 2620

SSP 3109SSP 3302

SSP 3304

SSP 3521

SSP 5310

SSP 5509

SSP 6208

SSP 6407

SSP 6512

SSP 6606

SSP 7218

SSP 7336

SSP 7419

SSP 7511SSP 7703SSP 7720

SSP 8113

SSP 8201

SSP 8202

SSP 8206

SSP 8208

SSP 8213SSP 8224SSP 8228

SSP 8234

SSP 8235SSP 8313

SSP 8317

SSP 8330

SSP 8346

SSP 8408

SSP 8412

SSP 8416SSP 8523

SSP 8531

SSP 8607

SSP 8613

SSP 8615

SSP 8703 SSP 9504

Fig. 3. Principal component analysis biplot obtained after removal ofspurious variables. Samples from white seabream larvae fed the controldiet, diet Low or diet High are labelled with C, L and H, respectively.The first principal component (PC1) accounted for 56% of theobserved variance and the second (PC2) for 18%

C1

C2

C3

C4 L1 L2 L3 L4 H1

H2

H3

H4

SSP 2121SSP 0124SSP 6512SSP 2517SSP 6407SSP 3304SSP 1405SSP 1402SSP 3109SSP 1704SSP 3521SSP 1424SSP 8346SSP 8208SSP 7336SSP 8330SSP 1201SSP 7218SSP 8206SSP 8613SSP 8317SSP 7419SSP 3302SSP 8213SSP 8224SSP 8228SSP 1327SSP 2309SSP 5509SSP 5310SSP 8234SSP 8615SSP 8416SSP 8531SSP 2620SSP 7511SSP 8607SSP 8235SSP 8412SSP 8113SSP 8523SSP 6606SSP 7720SSP 8703SSP 7703SSP 8408SSP 6208SSP 8201SSP 8202SSP 8313SSP 9504SSP 5503

Fig. 4. Heatmap showing relative abundance of identified proteins forall samples. Spots were grouped using Euclidian distance by agglom-erative hierarchical clustering (complete linkage method). Dark shadesindicate a lower than average expression of protein spots and lightshades a higher than average expression. Samples from white seabreamlarvae fed the control diet, diet L or diet H are labelled with C, L andH, respectively. �SSP�s indicates protein spot number identification

Dietary protein: skeleton and proteome in larvae 481

Page 6

Table

4Protein

spotsidentified

byLC-M

S⁄M

SorMALDI-ToF

⁄ToFMS,usingpeptidefragmentfingerprinting(PFF)in

theoptionMS

⁄MSIonSearchfrom

thebioinform

atics

applicationMascot.ThePFFwas

madein

thenonredundantNCBInrdatabase

fortheActinopterygiitaxonomic

level

(P-value<

0.05)

SpotN.

Highestscore

hit

Glnumber

(species)

pl t

⁄Mwt

ple

⁄MWea

Coverage

(%)

PM

bBestpeptidematch:sequence,

chargestate,E-value

Combined

mowse

score

c

0124

Fast

skeletalmyosinlightchain

3gi|5852836(Sparusaurata)

4.3

⁄13.2

4.4

⁄17.0

45

6GTYDDYVEGLR,+

2,1.5E-04

400

1201

Apolipoprotein

A-IV4

gi|74096419(Takifugurubripes)

4.8

⁄28.5

4.6

⁄27.2

82

LDPYAQDLQAR,+1,2.6E-06

132

1327

14-3-3

protein

gi|46326988(Oncorhynchusmykiss)

4.7

⁄29.2

4.7

⁄28.9

28

2YLAEFATGNDR,+1,1.9E-02

76

1402

Tropomyosinalpha-1

chain

gi|60390740(Lizaaurata)

4.6

⁄40.4

4.7

⁄32.8

42

6TID

DLEDELYAQK,+2,3.4E-06

644

1405

Tropomyosinalpha-1

chain

gi|60390740(Lizaaurata)

4.7

⁄38.7

4.7

⁄32.8

40

6TID

DLEDELYAQK,+2,2.0E-06

554

1424

Tropomyosin

gi|295792268(Epinepheluscoioides)

4.7

⁄32.7

4.7

⁄41.1

56

3KLVIIEGDLER,+

1,3.3E-08

235

1704

Protein

disulphideisomerase

(sim

ilarto)

gi|47223959(Tetraodonnigroviridis)

4.6

⁄76.1

4.7

⁄55.2

20

5VDATEETELAQEFGVR,+

2,1.OE-05

486

2121

Cytoplasm

icbetaactin

gi|166202369(Oncorhynchuskisutch)

5.6

⁄11.2

5.0

⁄16.7

21

n.a.

n.a.

80(98.2%

)

2309

14-3-3

protein

gamma

gi|237769613(Thunnusorientalis)

4.9

⁄28.3

4.9

⁄31.9

18

1ELEAVCQDVLNLLDNFLIK

,+1,100%

122

2517

TypeII

keratinE3-likeprotein

gi|48476437(Sparusaurata)

4.9

⁄38.6

4.9

⁄39.4

58

1VDALQDEIN

FLR,+

1,1.0E-05

72

2620

ATPsynthase

subunitbeta,

mitochondrial

gi|198285477(Salm

osalar)

4.9

⁄52.9

5.0

⁄58.8

14

2IV

AVIG

AVVDVQFDEGLP,+1,1.6E-17

315

3109

ATPsynthase

subunitbeta,

mitochondrial

gi|198285477(Salm

osalar)

4.9

⁄52.9

5.1

⁄14.5

44

3IPVGPETLGR,+

1,3.2E-08

315

3302

Calbindin

2a

gi|41152295(Danio

reho)

5.0

⁄31.1

5.1

⁄30.5

16

2IE

MSELAQIL

PTEENFLLC,+

1,8.0E-04

152

3304

Proteasomesubunitalphatype-3

gi|229366168(Anoplopomafimbria)

5.2

⁄28.4

5.1

⁄33.9

81

AFELELSWVGEVTNGR,+

1,1.6E-10

98

3521

TypeII

keratinE3-likeprotein

gi|48476437(Sparusaurata)

4.9

⁄38.6

5.0

⁄40.8

63

3VDALQDEIN

FLR,+

1,7.1

E-05

167

5310

Skeletalalphaactin

gi|291167454(Cobitischoii)

5.2

⁄42.3

5.5

⁄29.7

48

1SYELPDGQVIT

IGNER,+

1,3.0E-02

103(9.0E-06)

5503

Skeletalalphaactin

gi|6653228(Sparusaurata)

5.3

⁄42.2

5.3

⁄43.0

49

2AVFPSIV

GRPR,+

1,4.8-04

136

5509

Alphaactin

gi|8489855(Salm

otrutta)

5.2

⁄41.9

5.5

⁄51.5

27

1AG

FAG

DDAPR,+

1,100%

190

6208

Smallheatshock

protein

gi|226439776(Epinepheluscoioides)

6.4

⁄27.1

5.7

⁄25.8

59

2WDTWSNSYR,+

1,1.0E-03

93

6407

Skeletalalpha-actin

type-2b

gi|30268609(Coryphaenoides

yaquinae)

5.7

⁄37.5

5.2

⁄42.2

42

13

DLYANNVLSGGTTMYPGIA

DR,

+2,2.2E-07

469

6512

Skeletalalphaactin

gi|291167454(Cobitischoii)

5.2

⁄42.3

5.6

⁄38.8

48

n.a.

n.a.

135(5.7E-09)

6606

Glutaminesynthetase

gi|18252824(Bostrychussinensis)

5.7

⁄41.3

5.7

⁄51.8

25

n.a.

n.a.

87(3.3E-04)

7218

ATPsynthase

subunitalpha,

mitochondrial

gi|213512628(Salm

osalar)

9.0

⁄57.2

6.1

⁄25.5

20

1TGAIV

DVPVGEELLGR,+1,1.9E-04

78

482 M. de Vareilles et al.

Page 7

Table

4Continued

SpotN.

Highestscore

hit

Glnumber

(species)

pl t

⁄Mwt

ple

⁄MWea

Coverage

(%)

PM

bBestpeptidematch:sequence,

chargestate,E-value

Combined

mowse

score

c

7336

Myosin-bindingprotein

H-like

gi|317418695(Dicentrarchuslabrax)

5.8

⁄57.8

6.1

⁄30.1

21

TGDWFTVLEHYHR,+

1,100%

97

7419

Proteasome,

subunitbetatype8

gi|315518855(Oryziascelebensis)

8.0

⁄31.0

6.2

⁄40.5

41

n.a.

n.a.

67(3.6E-02)

7511

Betatubulin

gi|10242162(Notothenia

coriiceps)

4.8

⁄50.2

6.0

⁄44.8

56

1FPGQLNADLR,+1,1.2E-05

71

7703

Heatshock

protein

9gi|28278640(Danio

reho)

7.0

⁄74.2

5.9

⁄73.7

18

4VLGQFTLVGIPPAPR,+1,3.2E-08

306

7720

Antiquitin

gi|188036012(Acanthopagrusschlegelii)

5.9

⁄55.1

6.1

⁄56.6

30

n.a.

n.a.

73(8.4E-03)

8113

Nucleosidediphosphate

kinase

gi|194500331(Sparusaurata)

6.4

⁄17.1

6.5

⁄16.8

40

2TFIA

IKPDGVQR,+1,3.5E-04

107

8201

BetaA2crystallin

gi|77024823(Dissostichusmawsoni)

6.3

⁄23.8

6.0

⁄23.7

48

9CEFMLECQNIM

ER,+2,2.3E-06

572

8202

GammaN2crystallin

gi|1222522581(Poecilia

reticulata)

6.3

⁄21.6

6.3

⁄24.1

22

3VFGDGAWVMYEEPNFR,+

1,100%

92

8206

Cofilin-2

gi|213515222(Salm

osalar)

6.6

⁄18.7

6.2

⁄15.8

36

1YGLYDATYETK,+1,8.6E-04

53

8208

BetaA1crystallin

gi|72535901(Dissostichusmawsoni)

6.1

⁄23.0

6.4

⁄25.5

25

n.a.

n.a.

88(99.8%)

8213

BetaA1-2

crystallin

gi|221048019(Epinepheluscoioides)

6.1

⁄24.9

6.5

⁄25.1

45

2NWGSHCQTPQIQ

SIR

,+

1,100%

172

8224

BetaA4crystallin

gi|77024825(Dissostichusmawsoni)

6.4

⁄23.0

6.6

⁄24.1

29

1IIVFDEECFQGR,+

1,100%

86

8228

BetaA1-2

crystallin

gi|221048019(Epinepheluscoioides)

6.1

⁄24.9

6.7

⁄25.5

50

1HSGDYQHWR,+1,7.5E-04

83

8234

BetaA4crystallin

gi|77024825(Dissostichusmawsoni)

6.8

⁄22.3

6.4

⁄23.7

50

6IIVFDEECFQGR,+

2,2.4E-07

410

8235

Cofilin-2

gi|229366360(Anoplopomafimbria)

6.8

⁄18.9

6.8

⁄18.8

15

1VTDEVIA

VFNDMK,+1,100%

82

8313

Proteasomesubunitalphatype1

gi|62079624(Oreochromismossambicus)

9.1

⁄22.8

6.4

⁄30.5

37

n.a.

n.a.

68(2.7E-02)

8317

BetaB1crystallin

gi|77024827(Dissostichusmawsoni)

6.4

⁄26.4

6.7

⁄30.1

28

2SIIVECGPFVAFEQTNFR,+1,3.2E-11

209

8330

Creatinekinase

isoform

agi|156972295(Hippoglossushippoglossus)

6.9

⁄27.2

6.9

⁄27.5

26

6GGDDLDPNYVLSSR,+

2,6.3E-06

387

8346

Muscle-typecreatinekinase

M1

gi|21694041(Oreochromismossambicus)

6.6

⁄27.6

7.0

⁄43.3

16

5TFLVWVNEEDHLR,+

2,4.4E-06

339

8408

Glyceraldehyde3-phosphate

dehydrogenase

gi|15146358(Pagrusmajor)

6.6

⁄39.5

6.4

⁄36.4

43

9VPVADVSVVDLTCR,+2,6.0E-07

624

8412

Creatinekinase

M2

gi|4027927(Cyprinuscarpio)

6.2

⁄42.9

6.7

⁄36.9

91

TFLVWVNEEDHLR,+

1,99.9%

77

8416

Creatinekinase

isoform

agi|156972295(Hippoglossushippoglossus)

6.9

⁄27.5

6.5

⁄33.2

27

n.a.

n.a.

82(1.2E-03)

8523

Chain

A,crystalstructure

of

zebrafish

Ape

gi|162329921(Danio

rerid)

5.5

⁄32.3

6.6

⁄45.7

28

1IT

SWNVDGLR,+1,5.0E-02

35

Dietary protein: skeleton and proteome in larvae 483

Page 8

Discussion

Effect of dietary protein hydrolysates on larval development

It has been suggested that including PH in inert diets forlarviculture can improve performance and quality because the

dietary nitrogen source thus provided is more readily availablethan intact proteins to finfish larvae, which generally do nothave a fully developed and functional digestive system (see

Savoie et al., 2011; and references therein). Interestingly, ourresults showed for the first time that a 15% inclusion of largerpolypeptides (0.5–30 kDa) together with a 5% inclusion ofFAA and di- and tri-peptides (diet H) improved growth of

white seabream larvae at 15 DAH, whereas including only 5%larger polypeptides but 15% FAA, di- and tri-peptides (diet L)reduced the growth rate of larvae, compared to the control

larvae (approximately 5% larger polypeptides and 5% FAA,di- and tri-peptides). It could be that diet L presented an excessof FAA, di- and tri-peptides that saturated the peptide and AA

intestinal transport mechanisms, resulting in a reduced early-stage performance, as suggested by Cahu et al. (1999) andCarvalho et al. (2004), and also discussed in Tonheim et al.

(2005).The growth rates and survival observed in this study were

slightly lower than those previously reported for this species(Saavedra et al., 2009b, 2010b; Guerreiro et al., 2010) and this

is most probably due to the very early weaning of larvae ontomicrodiets. Substitution of live feed for inert diets in marinelarviculture has been difficult to achieve with success in terms

of survival and growth. White seabream weaning is usuallyperformed after 20 DAH, when the digestive tract develop-ment is complete and the stomach becomes functional (Cara

et al., 2003; Ortiz-Delgado et al., 2003; Guerreiro et al., 2010).In the present study, larvae were in contact with the microdietsfrom first feeding onwards so as to enable the study of theireffects on development as early as possible, but this may have

lead to an increase in the mortality rate, especially for smallerlarvae.

To date, few studies have focused on the larval skeletal

deformity issue in white seabream (Saavedra et al., 2009a,b,2010a,b), whose larval stages seem to be as affected as othersparids by this problem in farming conditions. The percentage

of deformed D. sargus larvae can reach up to 90% of thefarmed individuals depending on the developmental stages ofthe larvae (Saavedra et al., 2010a). In the present study, the

percentage of deformed individuals registered at 28 DAH,ranged between 60 and 69%, which was lower than valuesreported for white seabream larvae reared on live food(Saavedra et al., 2010a) and slightly lower than the incidence

of deformed larvae observed when fed on a microencapsulateddiet in the study of Saavedra et al. (2010b). In other species ofsparids, such as gilthead seabream and sharpsnout seabream,

the caudal complex, has been identified as the regions moreseverely affected by skeletal abnormalities during larvaldevelopment (Favaloro and Mazzola, 2000; Boglione et al.,

2001; Fernandez et al., 2008). In white seabream larvae fedeither live or microencapsulated diets, Saavedra et al.(2010a,b) also observed a higher occurrence of skeletal

deformities in the preurostyle region around 25–30 DAH.The incidence of skeletal deformities affecting the caudalvertebrae and caudal fin complex (D+G) ranged from 38.6 to41.8% of the deformed larvae, which was lower than the

48.5% previously recorded on hatchery-reared gilthead seab-ream larvae (Boglione et al., 2001) but higher than the 29%recently observed by Fernandez et al. (2008) at a latter

Table

4Continued

SpotN.

Highestscore

hit

Glnumber

(species)

pl t

⁄Mwt

ple

⁄MWea

Coverage

(%)

PM

bBestpeptidematch:sequence,

chargestate,E-value

Combined

mowse

score

c

8531

Muscle

typecreatinekinase

CKM1

gi|268315573(Platichthysstellatus)

6.2

⁄43.2

6.6

⁄44.5

34

1GFTLPPHNSR,1+

,1.6E-04

80

8607

Enolase-1

(alpha)

gi|37590349(Danio

rerio)

6.2

⁄47.4

6.4

⁄57.6

43

1IG

AEVYHNLK,+1,7.9E-05

91

8613

Enolase-2

(gamma)

gi|6624237(Lethenteronreissneri)

6.3

⁄43.3

6.7

⁄53.6

30

1EVIL

PVPAFNVIN

GGSHAGNK,+

1,99.2%

63

8615

Beta-enolase

(3)

gi|295792264(Epinepheluscoioides)

6.3

⁄47.8

6.8

⁄59.6

39

1GNPTVEVDLWTAK,+1,5.4E10-06

74

8703

Pre-m

RNA

processingfactor19

gi|291190276(Salm

osalar)

6.1

⁄54.8

6.4

⁄64.1

91

SLVFDQSGTYLAVGGSDI,+1,100%

111

9504

Fructose-bisphosphate

aldolase

Cgi|46849419(Acipenserbaerii)

5.4

⁄35.9

6.7

⁄42.6

21

ALQASALNAWR,+1,2.7E-06

77

aTheoreticalisoelectricalpoints

(pI t)andmolecularweights

(Mwt)basedonthebestresult�ssequence;experim

entalisoelectric

points

(pI e)andmolecularweights

(Mwe)estimatedfrom

thepositionofthe

spots

inthegels;Mw

isgiven

inkDa.

bNumber

ofpeptides

matched

(E-value<

0.05),

�n.a.�when

PFFgavenosignificantresults.

cA

non-probabilisticprotein

score,derived

from

theionsscore

(theMowse

score

andcorrespondingE-valuefrom

peptidemass

fingerprinting(PMF)isgiven

when

PFFgavenosignificantresult).Fordata

provided

bytheMass

Spectrometry

Laboratory,ASU,IT

QB,UNL,theconfidence

interval(%

)isgiven

insteadoftheE-value.

484 M. de Vareilles et al.

Page 9

developmental stage (60 DAH) in the control group of theirstudy, conducted on gilthead seabream fed with enriched live

preys. In our results it is remarkable that the cephalic part ofthe vertebral column of the larvae was more heavily affected byskeletal deformities than the caudal complex, representing

69.4–72.7% of the deformed larvae that presented at least onedeformity in this region. This pattern of deformities was notpreviously reported in D. sargus and can be due not to thetreatment itself but to intrinsic factors of the spawn, since a

similar distribution is observed in all treatment groups. Indusky grouper, an increasing occurrence of cephalic deformi-ties (including head and cephalic vertebrae) was observed when

increasing larval stocking densities (Boglione et al., 2009). Inour work the larval density corresponded to half the densityrecommended for commercial gilthead seabream hatchery

production (Moretti et al., 1999), and should not be respon-sible for the high incidence of cephalic deformities recorded.The cephalic area of the vertebral column of the larvae wasmostly affected by malformations in the neural arches and ⁄ orspines (A5), which were also the most commonly deformedstructures encountered along the vertebral column, accountingfor 83.3–100% of the deformities in the cephalic area. Severe

types of deformities such as lordosis or kyphosis were notrelevant in our study, with only one larvae showing lordosis.These deformities are commonly encountered in hatcheries

(Favaloro and Mazzola, 2000, 2003; Boglione et al., 2001;Lewis et al., 2004) and have been assumed to be related tophysicochemical parameters in rearing tanks such as hydro-

dynamics (Divanach et al., 1997) and temperature (Georgak-opoulou et al., 2010) but also with failure of swimbladderinflation (Chatain, 1994). Previous studies conducted onEuropean seabass (Dicentrarchus labrax) have shown that

the occurrence of deformed larvae can be gradually diminishedby incorporating graded levels of both commercial PH (CPSPG) and fish PH fraction composed mainly (75%) by di- and tri-

peptides (Zambonino Infante et al., 1997; Cahu et al., 1999).In the present study, we did not observe any particularbeneficial effect on the incidence of deformed white seabream

larvae at 28 DAH, when comparing the PH fractions ofdifferent molecular weight.

Whole body proteome response

Feeding early white seabream larvae with different composi-tions of fish PH clearly affected the expression profile of

soluble, abundant proteins with molecular weight between 10and 150 kDa and a pI value between 4 and 7. Interestingly, wesaw that the inclusion of higher fish PH content in the larval

diet has a more pronounced effect on whole body proteomeprofile at 15 DAH than the different size of the hydrolysatesincluded, the proteome of the control diet-fed larvae being

clearly more dissimilar to those of larvae fed the experimentalPH (diets L and H).A large number of protein spots identified corresponded to

cytoskeletal proteins, that play key roles in the maintenance of

cell architecture, cell motility, proliferation, differentiation andorganelle transport. Alpha actin, a major constituent of muscletissue, corresponded to almost a third of these protein spots

and seemed to be both up-and down-regulated across dietarytreatment. This can probably be attributed to the existence ofmultiple isoforms and ⁄ or post-translational modifications

whose occurrence seems to be regulated by numerous andcomplex mechanisms. Myosin light chain isoforms have alsobeen shown to be closely regulated, particularly during fish

development (Silva et al., 2010 and others therein) andinterestingly, a similar isoform, along with beta actin, were

also shown to be down-regulated in cod larvae fed with fish PH(Sveinsdottir and Gudmundsdottir, 2010). Type II keratinsexpression has also been shown to be developmental stage-

specific in some telesost larvae, being down-regulated in laterstages of development (Sarropoulou et al., 2005; Sveinsdottiret al., 2008; Infante et al., 2011) and thus could indicate abetter development of larvae fed diet H, in which they were

under-expressed, which is consistent with the growth results.Of the remaining cytoskeletal proteins, cofilin-2 and beta

tubulin were over-expressed in groups L and H relative to the

control group, possibly indicating increased cytoskeletal reor-ganisation in larvae fed a higher dietary content of PH.Consistent with, but not restricted to, enhanced cytoskeletal

dynamics, all proteins identified as belonging to energymetabolic processes were over-expressed in the larvae fedexperimental fish PH compared to control larvae. In the caseof larvae fed diet H, the increase in muscle creatine kinase

expression could simply be due to the high energy require-ments of an increased growth rate, as seen for example byGuderley et al. (2001) in threespine stickleback. The over-

expression of muscle creatine kinase, mitochondrial ATPsynthase subunits and proteins belonging to the glyco-lytic ⁄ gluconeogenic pathways in larvae fed diet L may reflect

a more important energy allocation to primary metabolismand higher protein catabolism. Studies show that dietary AAimbalances increase protein catabolism (Aragao et al., 2004)

and indeed, regulation of protein synthesis is a promisingmeans to limit energy expenditures under unfavourable feedingconditions (Salem et al., 2007). If we assume that diet Lpresented an excess of FAA, di- and tri-peptides, thereby

impairing normal dietary nitrogen absorption and utilisation,then larvae from this group might have been suffering fromenhanced protein catabolism to meet basal energy require-

ments. In agreement with this are the over-expression ofglutamine synthetase and antiquitin in this group. Glutaminesynthetase catalyses the ATP-dependent conversion of gluta-

mate and ammonium to glutamine, playing a key role innitrogen metabolism and is critical for the detoxificationprocess of ammonia, a common product of protein catabolismand gluconeogenesis (Essex-Fraser et al., 2005; and others

therein). Antiquitin, or aldehyde dehydrogenase 7, is amultifunctional enzyme that protects cells from oxidativestress by metabolising a number of lipid peroxidation-derived

aldehydes, has been shown to act as an osmo-protectant and isdirectly involved in lysine catabolism (Tang et al., 2008,references therein), thus its increased abundance could be

pointing to a more active AA metabolism.The protein pre-mRNA splicing factor 19, a conserved

eukaryotic RNA splicing factor and member of WD40 repeat

family of E3 ubiquitin ligases was over-expressed in group L,along with the proteasome subunit alpha type 1, which couldalso be pointing to an increased activation of the proteasomedegradation pathway and thus higher protein catabolic activ-

ity. However, other subunits of the 20S core particle structureof the proteasome were observed with different expressionpatterns, which is not surprising given the variety of interven-

tion points of the proteasome, suggesting that the role of eachsubunit should be further investigated.The protein disulfide isomerase, essential for correct folding,

stability and ⁄ or multimerisation of many proteins, was alsoover-expressed in group L, relative to group H, possiblyindicating a higher demand from the unfolded protein

Dietary protein: skeleton and proteome in larvae 485

Page 10

response pathway in the ER lumen of these larvae. However,other proteins with chaperone activity were also over-

expressed in both groups H and L, relative to the controlgroup, such as the small heat shock protein and HSP9. Indeed,quite a number of proteins (nuclear Ape, NDK, 14-3-3,

calbindin) implicated in cell-cycle progression, stress response,survival pathways and metabolic regulatory pathways, wereover-expressed in both groups L and H relative to the controlgroup and shows the significant impact that changing the

dietary fraction of fish PH has on whole body proteome ofwhite seabream larvae.Interestingly, apolipoprotein A-IV4 was over-expressed in

group H. Studies show that this protein may play a unique rolein integrating appetite regulation, intestinal lipid absorptionand energy storage in mammals (Simon et al., 2011). Assuming

Apo A-IV4 has a similar role in developing white seabream, weinterpret the signs of increased expression of this protein in thegroup which displayed higher weight gain as a sign that thedietary PH molecular size fraction and percentage in diet H

might have enhanced appetite regulation and triglycerideabsorption in these larvae through an induced increase inApo A-IV4 expression.

Finally, a number of crystallin isoforms and ⁄ or PTMs wereidentified, under-expressed in the control group compared tothose fed the experimental PH fractions. Crystallins were

initially characterised as structural proteins of the vertebratelens, however, it is now believed that many crystallins mayplay a number of non-lens biology-related roles (Weadick and

Chang, 2009, references therein) and further research isnecessary for establishing the specific roles played by crystal-lins.This study suggested that the molecular weight of the

dietary PH fraction had a minor impact on skeletal defor-mities appearance in white seabream larvae. However, itcannot be excluded that other factors than the ones tested

may have masked the effect of the diet on the skeletogenesisof white seabream, particularly, the early weaning schemewhich might have induced a slightly lower growth rate and

thus observations at a later larval developmental stage mayhave allowed detection of significant effects. Still, dietarytreatments affected growth performance and had a strongimpact on larvae whole body proteome. To our best

knowledge this is the first study using comparative proteo-mics with such a high number of proteins identified, in teleostlarval proteome. The proteins identified were involved in

diverse processes such as, cytoskeletal dynamics, energy andcarbohydrate metabolism, lipoprotein metabolism, AAmetabolism, nucleotide metabolism, protein chaperoning

and degradation, and signal transduction. This provides agood starting point to further investigate specific roles ofdietary PH on development of teleost larvae.

Acknowledgements

This study was partially supported by the projects SAARGO –

POCI ⁄ MAR ⁄ 6123 ⁄ 2004 and HYDRAA – PTDC ⁄ MAR ⁄71685 ⁄ 2006, granted by Fundacao para a Ciencia e Tecno-logia, Portugal. NR, M de V and TSS acknowledge financial

support from Fundacao para a Ciencia e Tecnologia, Portugal,through grants SFRH ⁄ BDP ⁄ 34888 ⁄ 2007, SFRH ⁄BD ⁄ 40698 ⁄ 2007 and SFRH ⁄ BD ⁄ 41392 ⁄ 2007 respectively.

The authors would like to thank Monica Mai and RicardoAlves for their useful help during sampling and Dr. EckhardWitten for useful suggestions for table presentation.

References

Alves, R. N.; Cordeiro, O.; Silva, T. S.; Richard, N.; de Vareilles, M.;Marino, G.; Di Marco, P.; Rodrigues, P. M.; Conceicao, L. E. C.,2010: Metabolic indicators of chronic stress in gilthead seabream(Sparus aurata) using comparative proteomics. Aquaculture 299,

57–66.Aragao, C.; Conceicao, L. E. C.; Fyhn, H.-J.; Dinis, M. T., 2004:

Estimated amino acid requirements during early ontogeny infish with different life-styles: gilthead seabream (Sparus aurata)and senegalese sole (Solea senegalensis). Aquaculture 242, 589–605.

Boglione, C.; Gagliardi, F.; Scardi, M.; Cataudella, S., 2001: Skeletaldescriptors and quality assessment in larvae and post-larvae ofwild-caught and hatchery-reared gilthead sea bream (Sparusaurata L. 1758). Aquaculture 192, 1–22.

Boglione, C.; Marino, G.; Giganti, M.; Longobardi, A.; De Marzi, P.;Cataudella, S., 2009: Skeletal anomalies in dusky grouperEpinephelus marginatus (Lowe 1834) juveniles reared with differ-ent methodologies and larval densities. Aquaculture 291, 48–60.

Cahu, C. L.; Zambonino Infante, J. L.; Quazuguel, P.; Le Gall, M. M.,1999: Protein hydrolysate vs. fish meal in compound diets for 10-day old sea bass Dicentrarchus labrax larvae. Aquaculture 171,

109–119.Cara, J. B.; Moyano, F. J.; Cardenas, S.; Fernandez-Dias, C.; Yufera,

M., 2003: Assessment of digestive enzyme activities during larvaldevelopment of white bream. J. Fish Biol. 63, 48–58.

Carvalho, A. P.; Sa, R.; Oliva-Teles, A.; Bergot, P., 2004: Solubilityand peptide profile affect the utilization of dietary protein bycommon carp (Cyprinus carpio) during early larval stages.Aquaculture 234, 319–333.

Chatain, B., 1994: Abnormal swimbladder development and lordosisin sea bass (D. labrax) and sea bream (S. aurata). Aquaculture119, 371–379.

Conceicao, L. E. C.; van der Meeren, T.; Verreth, J. A. J.; Evjen, M.S.; Houlihan, D. F.; Fyhn, H. J., 1997: Amino acid metabolismand protein turnover in larval turbot (Scophthalmus maximus) fednatural zooplankton or Artemia. Mar. Biol. 129, 255–265.

Dabrowski, K.; Lee, K.-J.; Rinchard, J., 2003: The smallest vertebrate,teleost fish, can utilize synthethic dipeptide-based diets. J. Nutr.133, 4225–4229.

Divanach, P.; Papandroulakis, N.; Anastasiadis, P.; Koumoundouros,G.; Kentouri, M., 1997: Effect of water currents on the develop-ment of skeletal deformities in sea bass (Dicentrarchus labrax L.)with functional swimbladder during postlarval and nursery phase.Aquaculture 156, 145–155.

Essex-Fraser, P. A.; Steele, S. L.; Bernier, N. J.; Murrays, B. W.; DonStevens, E.; Wright, P. A., 2005: Expression of four glutaminesynthetase genes in the early stages of development of rainbowtrout (Oncoryhnchus mykiss) in relationship to nitrogen excretion.J. Biol. Chem. 280, 20268–20273.

Favaloro, E.; Mazzola, A., 2000: Meristic character analysis andskeletal anomalies during growth in reared sharpsnout seabream.Aquac. Int. 8, 417–430.

Favaloro, E.; Mazzola, A., 2003: Meristic variation and skeletalanomalies of wild and reared sharpsnout seabream juveniles(Diplodus puntazzo, Cetti 1777) off coastal Sicily, MediterraneanSea. Aquac. Res. 34, 575–579.

Favaloro, E.; Mazzola, A., 2006: Meristic character counts andincidence of skeletal anomalies in the wild Diplodus puntazzo(Cetti, 1777) of an area of the south-eastern Mediterranean Sea.Fish Physiol. Biochem. 32, 159–166.

Fernandez, I.; Hontoria, F.; Ortiz-Delgado, J. B.; Kotzamanis, Y.;Estevez, A.; Zambonino-Infante, J. L.; Gisbert, E., 2008: Larvalperformance and skeletal deformities in farmed gilthead seabream (Sparus aurata) fed with graded levels of Vitamin Aenriched rotifers (Brachionus plicatilis). Aquaculture 283, 102–115.

Finn, R. N.; Rønnestad, I.; van der Meeren, T.; Fyhn, H. J., 2002: Fueland metabolic scaling during the early life stages of Atlantic codGadus morhua. Mar. Ecol. Prog. Ser. 243, 217–234.

Gavaia, P. J.; Sarasquete, M. C.; Cancela, M. L., 2000: Detection ofmineralized structures in very early stages of development ofmarine teleostei using a modified Alcian blue-Alizarin red doublestaining technique for bone and cartilage. Biotech. Histochem. 75,79–84.

Georgakopoulou, E.; Katharios, P.; Divanach, P.; Koumoundouros,G., 2010: Effect of temperature on the development of skeletal

486 M. de Vareilles et al.

Page 11

deformities in Gilthead seabream (Sparus aurata Linnaeus, 1758).Aquaculture, 308, 13–19.

Gisbert, E., 2010: Protein hydrolysates in larval fish nutrition. Yeast,pig blood hydrolysates substitute for fishmeal in study. Globalaquaculture advocate 13, 73–74.

Gomez-Requeni, P.; de Vareilles, M.; Kousoulaki, K.; Jordal, A. O.;Conceicao, L. E. C.; Rønnestad, I., 2011: Whole body proteomeresponse to a dietary lysine imbalance in zebrafish (Danio rerio).Comp. Biochem. Physiol. Part D Genomics Proteomics 6, 178–186.

Guderley, H.; Leroy, P. H.; Gagne, A., 2001: Thermal acclimation,growth and burst-swimming of Threespine Stickleback: enzymaticcorrelates and influence of photoperiod. Physiol. Biochem. Zool.74, 66–67.

Guerreiro, I.; de Vareilles, M.; Pousao-Ferreira, P.; Rodrigues, V.;Dinis, M. T.; Ribeiro, L., 2010: Effect of age-at-weaning ondigestive capacity of white seabream (Diplodus sargus). Aquacul-ture 300, 194–205.

Infante, C.; Ponce, M.; Asensio, E.; Zerolo, R.; Manchado, M., 2011:Molecular characterization of a novel type II keratin gene(sseKer3) in the Senegalese sole (Solea senegalensis): differentialexpression of keratin genes by salinity. Comp. Biochem. Physiol.B, Biochem. Mol. Biol. 160, 15–23.

Kolkovski, S.; Tandler, A., 2000: The use of squid protein hydrolysateas a protein source in microdiets for gilthead seabream Sparusaurata larvae. Aquacult. Nutr. 6, 11–15.

Koumoundouros, G., 2010: Morpho-anatomical abnormalities inMediterranean marine aquaculture. In: Recent advances inaquaculture research. G. Koumoundouros (Ed.). TransworldResearch Network, Kerala, India, pp. 125–148.

Kvale, A.; Harboe, T.; Mangor-Jensen, A.; Hamre, K., 2009: Effects ofprotein hydrolysate in weaning diets for Atlantic cod (Gadusmorhua L.) and Atlantic halibut (Hippoglossus hippoglossus L.).Aquacult. Nutr. 15, 218–227.

Lewis, L. M.; Lall, S. P.; Witten, P. E., 2004: Morphologicaldescriptions of the early stages of spine and vertebral developmentin hatchery-reared larval and juvenile Atlantic halibut (Hippog-lossus hippoglossus). Aquaculture 241, 47–59.

Moretti, A.; Pedini Fernandez-Criado, M.; Cittolon, G.; Guidastri, R.,1999: Manual on hatchery production of seabass and giltheadseabream, Vol. 1. FAO, Rome, Italy, 194 pp.

Murray, H. M.; Lall, S. P.; Rajaselvam, R.; Boutilier, L. A.; Flight,R. M.; Blanchard, B.; Colombo, S.; Mohindra, V.; Yufera, M.;Douglas, S. E., 2010: Effect of early introduction of microen-capsulated diet to larval Atlantic halibut, Hippoglossus hippog-lossus L. assessed by microarray analysis. Mar. Biotechnol. 12,214–229.

Ortiz-Delgado, J. B.; Darias, M. J.; Canavate, J. P.; Yufera, M.;Sarasquete, C., 2003: Organogenesis of the digestive tract in thewhite seabream Diplodus sargus. Histological and histochemicalapproaches. Histol. Histopathol. 18, 1141–1154.

Pousao-Ferreira, P.; Dores, E.; Morais, S.; Narciso, L., 2001: Lipidrequirements of the white seabream (Diplodus sargus) larvae. In:Larvi 2001 – Fish and Shellfish Larviculture Symposium. SpecialPublication, No. 30. European Aquaculture Society, Oostende,Belgium, 87 pp.

R Development Core Team, 2011: R: a language and environment forstatistical computing. R Foundation for Statistical Computing,Vienna, Austria. ISBN 3-900051-07-0, URL http://www.R-project.org/ (accessed on January 05, 2011).

Rønnestad, I.; Tonheim, S. K.; Fyhn, H. J.; Rojas-Garcıa, C. R.;Kamisaka, Y.; Koven, W.; Finn, R. N.; Terjesen, B. F.; Barr, Y.;Conceicao, L. E. C., 2003: The supply of amino acids during earlyfeeding stages of marine larvae: a review of recent findings.Aquaculture 227, 147–164.

Saavedra, M.; Conceicao, L. E. C.; Pousao-Ferreira, P.; Dinis, M. T.,2006: Amino acid profiles of Diplodus sargus (L., 1758) larvae:implications for feed formulation. Aquaculture 261, 587–593.

Saavedra, M.; Pousao-Ferreira, P.; Yufera, M.; Dinis, M. T.;Conceicao, L. E. C., 2009a: A balanced amino acid diet improvesDiplodus sargus larval quality and reduces nitrogen excretion.Aquacult. Nutr. 15, 517–524.

Saavedra, M.; Barr, Y.; Pousao-Ferreira, P.; Helland, S.; Yufera, M.;Dinis, M. T.; Conceicao, L. E. C., 2009b: Supplementation of

tryptophan and lysine in Diplodus sargus larval diet: effects ongrowth and skeletal deformities. Aquac. Res. 40, 1191–1201.

Saavedra, M.; Nicolau, L.; Pousao-Ferreira, P., 2010a: Developmentof deformities at the vertebral column in Diplodus sargus (L.,1758) early larval stages. Aquac. Res. 41, 1054–1063.

Saavedra, M.; Conceicao, L. E. C.; Barr, Y.; Helland, S.; Pousao-Ferreira, P.; Yufera, M.; Dinis, M. T., 2010b: Tyrosine andphenylalanine supplementation on Diplodus sargus larvae: effecton growth and quality. Aquac. Res. 41, 1523–1532.

Salem, M.; Silverstein, J.; Rexroad, C., III; Yao, J., 2007: Effect ofstarvation on global gene expression and proteolysis in rainbowtrout (Oncorhynchus mykiss). BMC Genomics 8, 328.

Santos, M. N.; Lino, P.; Pousao-Ferreira, P.; Monteiro, C. C., 2006:Preliminary results of hatchery-reared seabreams released atartificial reefs of the Algarve coast (Southern Portugal): a pilotstudy. Bull. Mar. Sci. 78, 177–186.

Sarropoulou, E.; Kotoulas, G.; Power, D. M.; Geisler, R., 2005:Gene expression profiling of gilthead seabream during earlydevelopment and detection of stress-related genes by applicationof cDNA microarray technology. Physiol. Genomics 23, 182–191.

Savoie, A.; Le Francois, N. R.; Lamarre, S. G.; Blier, P. U.;Beaulieu, L.; Cahu, C., 2011: Dietary protein hydrolysates andtrypsin inhibitor effects on digestive capacities and performancesduring early-stages of spotted wolffish: suggested mechanisms.Comp. Biochem. Physiol., Part A Mol. Integr. Physiol. 158, 525–530.

Silva, P.; Power, D. M.; Valente, L. M. P.; Silva, N.; Monteira, R. A.F.; Rocha, E., 2010: Expression of the myosin light chains 1, 2 and3 in the muscle of blackspot seabream (Pagellus bogaraveoBrunnich), during development. Fish Physiol. Biochem. 36,

1125–1132.Simon, T.; Cook, V. R.; Rao, A.; Weinberg, R. B., 2011: The impact of

murine intestinal apolipoprotein A-IV expression on regional lipidabsorption, gene expression, and growth. J. Lipid Res. 52, 1984–1994.

Sveinsdottir, H.; Gudmundsdottir, A., 2010: Proteome profile com-parison of two differently fed groups of Atlantic cod (Gadusmorhua) larvae. Aquacult. Nutr. 16, 662–670.

Sveinsdottir, H.; Vilhemsson, O.; Gudmundsdottir, A., 2008: Prote-ome analysis of abundant proteins in two age groups of earlyAtlantic cod (Gadus morhua) larvae. Comp. Biochem. Physiol.Part D Genomics Proteomics 3, 243–250.

Tang, W. K.; Chan, C. B.; Wong, K. B.; Lam, Y. M.; Cha, S. S.;Cheng, C. H. K.; Fong, W. P., 2008: The crystal structure ofseabream antiquitin reveals the structural basis of its substratespecificity. FEBS Lett. 582, 3090–3096.

Tonheim, S. K.; Espe, M.; Hamre, K.; Rønnestad, I., 2005: Pre-hydrolysis improves utilization of dietary protein in the larvalteleost Atlantic halibut (Hippoglossus hippoglossus L.). J. Exp.Mar. Bio. Ecol. 321, 19–34.

Weadick, C. J.; Chang, B. S. W., 2009: Molecular evolution of thebetagamma lens crystallin superfamily: evidence for a restrainedancestral function in gammaN crystallins? Mol. Biol. Evol. 26,1127–1142.

Yufera, M.; Pascual, E.; Fernandez-Dıaz, C., 1999: A highly efficientmicroencapsulated food for rearing early larvae of marine fish.Aquaculture 177, 249–256.

Yufera, M.; Fernandez-Dıaz, C.; Pascual, E., 2005: Food micropar-ticles for larval fish prepared by internal gelation. Aquaculture248, 253–262.

Zambonino Infante, J. L.; Cahu, C. L.; Peres, A., 1997: Partialsubstitution of di- and tripeptides for native proteins in sea bassdiet improves Dicentrarchus labrax larval development. J. Nutr.127, 608–614.

Zambonino Infante, J. L.; Cahu, C.; Villeneuve, L.; Gisbert, E., 2005:Nutrition, development and morphogenesis in fish larvae: somerecent developments. Aqua Feeds: Formulation & Beyond 2, 3–5.

Author�s address: Mahaut de Vareilles, CIMAR ⁄ CCMAR, Univer-sidade do Algarve, Campus de Gambelas, 8000-139Faro, Portugal.E-mail: [email protected]

Dietary protein: skeleton and proteome in larvae 487