Aquaculture nutrition, tập 19, số 6, 2013

1,2 2 2 2,3 2 2 22 1 Animal Science College, South China Agricultural University, Guangzhou, China; 2 School of Life Science, Sun Yat-Sen University, Guangzhou, China;3 Key Laboratory of

1,2 2 2 1 1

Skretting Aquaculture Research Centre, Stavanger, Norway; 2 Laboratory of Food Process Engineering, WageningenUniversity, Wageningen, the Netherlands

This article focuses on understanding the role of vital

wheat gluten on the structural parameters of extruded fish

feed and its correlation to the physical and functional

properties Gluten–soy protein concentrate blends with five

gluten concentrations (0–200 g kg
1) were produced An

abrupt reduction in oil uptake was observed with the 200 g

gluten kg
1 blend Inclusion of gluten from 100 to

200 g kg
1 resulted in unacceptable product properties

Sinking of feed pellets with 0 and 50 g gluten kg
1 was

100%, whereas only 36% of pellets with 200 g gluten kg
1

sank We suspect that this is due to a relationship between

morphological structure and oil impregnation during

coat-ing of feeds The addition of gluten at 200 g kg
1 gave a

smoother and non-porous outer surface Pellets without

gluten had a larger number of cells that were smaller than

200lm (P < 0.05) compared with pellets with 100 and

200 g gluten kg
1 More spherical cell shapes (P< 0.01)

and a compact structure were favoured in the presence of

gluten The closed porosity increased (P< 0.05), whereas

interconnectivity between pores decreased (P< 0.01), with

increasing gluten content from 0 to 200 g kg
1 The effects

of the addition of gluten are probably related to the

film-forming properties of gluten

KEY WORDS: extrusion, fish feed, microstructure, physical

quality, soy protein concentrate, wheat gluten

Received 8 July 2012; accepted 26 November 2012

Correspondence: Skretting ARC, PO Box 48, 4001 Stavanger, Norway.

E-mail: vukasin.draganovic@skretting.com

In an effort to increase formulation flexibility in the

pro-duction of modern salmonid feeds and to enhance the

sus-tainability of feeds by replacement of fish protein with

plant protein, vital wheat gluten has been shown to havehigh potential as a feed ingredient (Gatlin et al 2007).Among plant proteins, soy protein concentrate (SPC) isstill the primary alternative ingredient to fish meal due toits availability and competitive prices, but the high concen-tration of carbohydrates in SPC remains a concern (Gatlin

et al 2007) In this respect, vital wheat gluten may havegood potential due to its high protein content and nutrientdigestibility (Sugiura et al 1998; Robaina et al 1999), itslower level of indigestible fibres and absence of anti-nutri-tional factors Compared with fish meal, wheat gluten islow in methionine and especially low in lysine, whereas it ishigher in cysteine content (Allan et al 2000) It has beenreported previously that in salmonids, supplementationwith lysine (Davies et al 1997; Cheng et al 2003) or acombination of lysine and methionine (Pfeffer & Henrich-freise 1994) is required for diets containing wheat gluten tomaintain fish growth Wheat gluten is also extensively used

in food applications due to its functionality (Day et al.2006) and availability in large quantities (Domenek et al.2004)

Feed pellets obtained after extrusion should have a defined porosity that allows sufficient oil absorption capac-ity leading to specific sinking rates and durability (Glencross

well-et al.2010) In a commercial fish feed manufacturing tion, incorporation of gluten was shown to significantlyinfluence the physical properties of feed, such as oil infusion.For example, in high-fat feeds (>300 g kg
1added fat), glu-ten can be used only in small amounts (<100 g kg
1)because of the negative effect on this and other related tech-nical features of the feed The underlying mechanism affect-ing the physical properties of these extrudates is as yetunclear

opera-A quantitative description of the microstructure of theextrudate helps in understanding its mechanical properties(Robin et al 2010) The microstructural features of cellularproducts, such as average cell size, cell size distribution, cell

.

Aquaculture Nutrition

wall thickness, control product attributes such as the void

fraction and the interconnectivity of the cells (Gibson &

Ashby 1997) The infusion of liquids such as oil strongly

depends on cell size distribution, the degree of

interconnec-tivity between the cells and the average cell wall thickness

(Trater et al 2005) The precise influence of using wheat

gluten for partial or complete replacement of fish meal on

the morphology has not been described yet

Several techniques were used to study the microstructure

of expanded extrudates, such as scanning electron

micros-copy (SEM) (Warburton et al 1992) and light microsmicros-copy

(Chanvrier et al 2007) followed by digital imaging

(Sto-jceska et al 2008) However, these techniques only gave

information about a surface or a fracture plane X-ray

mic-rotomography (XMT) provides a non-invasive means of

assessing the morphology in three dimensions To the best

of our knowledge, this technique has not yet been used to

study the impact of plant proteins on the morphology of

fish feed extrudate

The specific objective of this study was to determine the

links between the physical and microstructural

characteris-tics of extruded feeds based on gluten–SPC blends andrelate these to the properties of the respective components

The microstructure of the pellets was altered through sion of different levels of wheat gluten mainly

inclu-All diets were processed at Skretting ARC TechnologyPlant (Stavanger, Norway) The formulations consisted of

a commercial salmon grower diet (0 g gluten kg
1) andfour experimental diets containing gluten, in which SPCwas partly replaced to give diets containing 50, 100, 150and 200 g gluten kg
1(50, 100, 150 and 200 g gluten kg
1diets, respectively) Commercially available SPC (Imcosoy)and gluten (vital wheat gluten powder) were obtained fromImcopa (Imcopa SA, Araucaria, Brazil) and Cargill (CargillGermany GmbH, Barby, Germany), respectively Table 1gives the formulations and chemical composition of thediets The chemical composition of the raw materials is

Table 1 Formulation and chemical composition of the dry feed mixes

2

crude fibre, 25; ash, 18.

3

fibre, 21; ash, 30.

142; starch, 5; crude fibre, 4; ash, 145.

crude fibre, 182; ash, 72.

.

given in the footnotes of Table 1 Chemical analysis of the

diets (Table 1) revealed that the inclusion of gluten led to a

lower level of crude fibre and higher level of starch and fat

However, no notable differences in the content of crude

protein and ash were observed between the feeds (Table 1)

The dry ingredients were premixed in a vertical mixer

(custom designed; Skretting ARC, Stavanger, Norway) and

ground in a Dinnissen 30 kW hammer mill (Dinnissen,

Sevenum, the Netherlands), with a screen size of 1.0 mm

Subsequently, the ingredients were mixed in a Dinnissen

500LTR horizontal ribbon mixer (Dinnissen) for 7 min

The feed mash was conditioned in a differential diameter

conditioner (DDC 2; Wenger Mfg Co., Sabetha, KS,

USA) and extruded in a Wenger TX-57 twin-screw

extru-der The barrel of the extruder was 57 mm in diameter and

the length-to-diameter ratio (L/D) was 17.5:1 The screw

configuration was composed of a series of intermeshing

feed screws (FS), a forwarding kneading block (FK) and

reversing kneading blocks (RK) arranged according to the

defined barrel diameters (D) such that the overall

configu-ration from the drive end was: 5D FS, 1D FK, 8D FS,

0.5D RK, 1D FS, 0.5D RK, 1.5D FS: to the die The

extru-der barrel consisted of four head sections, with each sectionjacketed to permit either steam heating (sections 1–4) orwater cooling (sections 2–4) Temperature control of thesecond, third and fourth section is achieved by balancingthe heating and cooling power input

The ingredients were processed as described, yielding lets with a diameter of 8.7 mm and a length of approxi-mately 10.0 mm The feed was dried in a Wenger Series IIIhorizontal 3-zones dryer (Wenger Mfg Co.) to approxi-mately 920 g kg
1 dry matter The allotted oil component

pel-of each diet was vacuum infused to the pellets in a Forberg6-l vacuum coater (pilot-scale coater) (Forberg®, Larvik,Norway) The mixtures were extruded using the parameterspresented in Table 2 Slight adjustments were made to thescrew speed and barrel temperature to obtain similarexpansion for each composition The knife rotation speedwas adjusted according to the specified length of the pel-lets Other extruder operating conditions were constant forall the feeds produced After at least 10 min of running,discrete samples of pellets (n= 4 9 1000 g) were collectedevery 15 min to create a repeated measures assessment ofeach diet

Table 2 Extruder and dryer processing parameters during feed production

Extrusion

Steam added to preconditioner

The specific mechanical energy (SME) and specific thermal

energy (STE) were read directly from the control panel of

the extruder The data were collected by APIS real-time

Pro-cess Explorer (version 4.1; Prediktor, Fredrikstad, Norway)

The values presented are the means of 10 measurements

taken at equal time intervals throughout the production run

Chemical analyses of the dry matter, protein, fat and ash

were carried out by Skretting ARC laboratory (an

accred-ited analytical service provider) The starch and crude fibre

analysis was carried out by Masterlab (Boxmeer, the

Neth-erlands) The dry matter was calculated by gravimetric

analysis after oven drying at 105°C for 18 h The protein

levels were calculated by determining the total nitrogen

content using the Kjeltec 2400 Auto System, based on

N 9 6.25 The fat concentration was measured by Maran

Ultra NMR nuclear magnetic resonance (Resonance

Instru-ments Ltd, Witney, UK) NMR measures the number of

hydrogen protons of some character in a sample This is

carried out by letting the protons induce a current in a

closed coil The sample was placed in an external magnetic

field, which aligns the protons, creating a net magnetic field

An RF pulse is applied to the sample, creating a dynamic

change in the field This induces a current in a coil

sur-rounding the sample The induced current is then related to

the fat content by a simple two-point calibration The gross

ash content was determined gravimetrically as the mass

remaining after combustion of a sample in a muffle furnace

at 550°C for 17 h Starch was analysed using an enzymatic

method described by McCleary et al (1994) The crude fibre

content was determined as the loss in mass resulting from

ashing of the dried residue obtained after acid and alkaline

digestion of the sample according to ISO 6865 (ISO 2000)

Digital imaging The pellets were photographed using a

zoom digital camera (Canon IXUS 210, Canon Inc.,

Head-quarters, Tokyo, Japan) with 14.1 effective megapixels for

high-definition pixel images

Light microscopy analysis Light microscopy was

per-formed using a Stereo Discovery V12 stereomicroscope

(Carl Zeiss, Goettingen, Germany) Pellets coated

(impreg-nated) with oil were cut in half along the cross-section with

a razor blade The pictures of the samples were taken with

the cutting side facing up

Scanning electron microscopy (SEM) Vacuum-dried ples were cut with a razor blade and glued onto a sampleholder using carbon adhesive tabs (Electron MicroscopySciences, Hatfield, PA, USA) or Leit-C (Neubauer Chemik-alien, M€unster, Germany), air dried for 2 h and subsequentlystored overnight under vacuum before analyses The sampleswere sputter coated with 20 nm of iridium in SCD 500 (Le-ica, Vienna, Austria) Samples were analysed at 2 kV atroom temperature in a Magellan 400 field emission SEM(FEI company, Eindhoven, the Netherlands) The imageswere digitally recorded The raw micrographs of the pelletsurfaces were reconstructed using Visiopharm image analysissoftware (VisioMorph– Visiopharm Integrator System®; Vi-siopharm, Hørsholm, Denmark) and analysed for porosity

sam-The program calculated the percentage of the image ing black pixels The porosity is reported as the area of pores

contain-as a percentage of the total area

Cryo-Scanning electron microscopy The frozen (
20 °C)sample was cut in half and glued onto a brass Leica sampleholder using carbon glue Leit-C (Neubauer Chemicalien),directly frozen in liquid nitrogen and simultaneously fittedinto the cryo-sample loading system (VCT 100) The Leicasample holder was transferred to a MED 020/VCT 100non-dedicated cryo-preparation system (Leica) on a samplestage at
93 °C The sample was freeze dried in this cryo-preparation chamber for 5 min at
93 °C and 1.3 9 10– 6

mbar to remove water vapour contamination from thesurface of the sample The sample was sputter coated with

a 15-nm layer of tungsten at the same temperature Thesample was then transferred into the Magellan 400 fieldemission SEM (FEI company) on the sample stage at

122 °C and 4 9 10
7mbar The analysis was performedwith a secondary electron voltage of 2 kV and probe cur-rent of 25 pA All images were recorded digitally

X-ray microtomography The 0, 100 and 200 g gluten kg
1feed pellets were scanned using a SkyScan 1172 desktop X-ray microtomography imaging system (SkyScan, Kontich,Belgium) with a pixel size of 5.4lm, operating at a voltage

of 50 kV and current of 160lA (to obtain optimumcontrast between the solid and gaseous phases) A 12-bit,11-megapixel, cooled CCD camera was used to collect theX-ray data The images were acquired with a rotation step

of 0.4° over a total rotation of 180° Image reconstructionwas accomplished using the Volumetric Reconstruction forMicro CT Instruments (SkyScan, Kontich, Belgium) soft-ware (Version 2.1, SkyScan, Kontich, Belgium) This recon-struction software uses a filtered back-projection algorithm

.

with a modified cone-beam reconstruction (Feldkamp et al.

1984) The measurements were limited to an appropriate

volume of interest (VOI), that is, a cylinder measuring

5.4 mm in diameter and 2.7 mm in height located in the

centre of each pellet The structural parameters were

calcu-lated by SkyScan CTAn software, version 1.5.13 (SkyScan,

Kontich, Belgium) after applying a global thresholding to

segment the solid and gaseous phases The porosity of the

pellets was calculated as the ratio of the volume of the

pores to the total volume of the pellet, where the pellet

vol-ume is equal to the VOI previously defined The structure

separation function yielded the cell size distribution, and

the thickness distribution function yielded the distribution

in thickness of the cell walls In addition, the open and

closed porosity are reported A closed pore in 3D is

charac-terized by a connected assemblage of space voxels that are

fully surrounded on all sides in 3D by solid voxels An

open pore is defined as any space located within a solid

object or between solid objects that have any connection in

3D to the space outside the object or objects Percent open

porosity is the volume of open pores as a percentage of the

total VOI volume The microstructure of the pellets was

also described in terms of average cell size, structure

sur-face per volume ratio, degree of anisotropy (DA) and

con-nectivity density The analyses were carried out on 3D

image All experiments were performed in four replicates

Specific density A volumetric displacement method using

glass beads with a diameter of 0.1 mm as a displacement

medium was used to determine the specific density of the

pellets including pores The method was originally

devel-oped by Hwang & Hayakawa (1980) The specific density

of the pellets was calculated using the following equation:

qs¼ Wpl

Vpg
ðWgs=qgsÞwhere qs is the specific density using the glass bead dis-

placement method (g L
1), Wpl is the pellet mass (g), Vpg

is the volume of the pellets and glass powder, Wgs is the

mass of glass beads displaced (g) andqgsis the specific

den-sity of the glass beads (g L
1) The values were obtained

from an average of three measurements

Oil absorption capacity The oil absorption capacity

(OAC) was measured according to a modified method of

Lin et al (1974) Approximately 0.5 g of the ground pelletsand 10.0 mL of rapeseed oil were added to a 15-ml conicalgraduated centrifuge tube The contents in the tube weremixed for 3 min with a vortex mixer to disperse the sampleinto the oil After a holding period of 30 min, the tube wascentrifuged for 25 min at 3050 g The separated oil wasthen removed with a pipette, and the tube was inverted for

25 min to drain the oil prior to reweighing The OAC wasexpressed as grams of rapeseed oil bound per gram ofground pellet Triplicate measurements were performed oneach sample

Maximum oil infusion into pellets Unlike the OAC, themaximum oil infusion test was performed under vacuumconditions, using the whole pellets The pellets (500 g) fromeach treatment were placed in a laboratory vacuum coater(custom designed; Skretting ARC), and the air was slowlyevacuated from the vacuum chamber down to a reducedpressure of 0.15 bar A mixture of heated (80°C) fish andrapeseed oil (50:50) was sprayed through a nozzle in anexcess amount (approximately 350 g) and thoroughly mixedwith the pellets throughout the coating cycle The vacuumchamber was then re-equilibrated to atmospheric pressure,and the oil was allowed to infuse into the feed The pelletswere then removed from the coater, and any excess oil wasremoved by placing the feed between two absorbent papertowels The final weight of the oil infused into the pellets wasthen recorded, and the relative oil uptake was calculated.The reported values were the average of three replicates

Fat leakage analysis Coated pellets from the previousanalysis were placed in a plastic bucket with blotting paper

in the bottom and stored at 40°C for 24 h The leakage offat was measured as the loss of fat from 100 g of feed

Holmen durability index This feed tester has recently beenintroduced in the fish feed industry It uses air to rapidlycirculate the feed and simulates a combination of mechani-cal and pneumatic stresses (Kaliyan & Vance Morey 2009).One hundred grams of coated product were placed into theNew Holmen Portable Pellet Tester (NHP 100; BorregaardLignotech, Sarpsborg, Norway), for 120 s The sampleswere then collected and weighed Three replicates weretested to calculate the Holmen durability index for eachdiet The chamber of the NHP 100 was cleaned, and the fil-ter paper was changed between each replicate The mea-surements are in line with the internal quality controlguidelines for coated pellets (Skretting AS, Stavanger, Nor-way), which ensures the industrial relevance of the results

Percentage of sinking pellets A 3000-ml cylinder was filled

with 40 g L
1 salt water One hundred pellets from each

treatment were randomly selected and individually dropped

into the cold water (6 °C) from 5 cm above the water

sur-face The pellets that did not sink within 3 min were given

a zero score, and percentage of sinking pellets was

calcu-lated The results were the average of three replicates

Water absorption index The water absorption index

(WAI) was determined by the method of Anderson et al

(1970) The pellets were first milled to a mean particle size

of approximately 270lm, determined by laser diffraction

analysis (Mastersizer 2000; Malvern Instruments Ltd.,

Mal-vern, UK) A 2.5 g sample was dispersed in 25 g of

dis-tilled water After stirring for 30 min, the dispersions were

rinsed into tarred centrifuge tubes, made up to 32.5 g and

then centrifuged at 3050 g for 10 min The supernatant

was decanted and the sediment was weighed The WAI was

calculated using the following equation: WAI= weight of

sediment/weight of dry solids All determinations were

con-ducted in triplicate

The results were submitted to one-way analysis of variance

(ANOVA) and a least significant difference test; a confidence

interval of 95% was used to compare the means Statistical

analyses were carried out using UNISTAT (Unistat

Com-puter Software Ltd, London, UK)

Figure 1 shows the effect of the inclusion of 200 g

glu-ten kg
1 on the appearance of the final product The

appearance of the products with 50, 100 and 150 g ten kg
1 was similar to the sample without gluten (resultsnot shown) When the 0 and 200 g gluten kg
1 sampleswere compared, a clear difference could be observed interms of oil uptake The surface of the sample with 0 g glu-ten kg
1was practically oil free, implying that most of theoil was captured inside the pellet Stereo light microscopywas used to evaluate the macroscopic morphology of thecross-sections of the 0 and 200 g gluten kg
1samples Fig-ure 2 clearly shows the region in the 200 g gluten kg
1sample that is not loaded with oil

glu-Table 3 lists the effects of gluten on the physical andfunctional properties of the feeds The density of the feedranged from 683 to 758 g L
1 Except for 150 g glu-ten kg
1, all the feeds had very similar density Althoughstatistically significant difference was observed, these differ-ences are not considered to be of technical relevance forthe effects explained in this study It is not completely clearwhy the feed with 150 g gluten kg
1 behaved differently

Obviously, this composition resulted in more expansioncompared with the other pellets The trends for the maxi-mum infusion of oil are known to closely follow density(Draganovic et al 2011) It was reported that the densityand maximum oil infusion had a correlation coefficient of

0.99 Here, except for 200 g gluten kg
1, we found thatthe maximum oil infusion trend was similar to that for thedensity for 0–150 g gluten kg
1.

It can be seen from Table 3 that the percentage of ing feed decreased (P< 0.01) with increasing gluten con-tent The lowest percentage sinking (35.7%) was obtained

sink-at 200 g gluten kg
1; the highest (100%) was observed at 0and 50 g gluten kg
1 and it was followed by 100 g glu-ten kg
1(94.7%)

Fat leakage varied from 3.5% to 6.8% It decreased asthe gluten level increased from 0 to 150 g kg
1; however,with 200 g gluten kg
1, a slightly higher value was again

Figure 1 Oil-coated pellet morphology for feed with 0 (a) and 200 g glu-

using pilot-scale coater.

.

found (Table 3) In general, the changes observed in fat

leakage also suggest differences in the morphology of the

products

Durability is typically manifested by retention of the

integrity of the feed during storage, transportation and

pneumatic feeding (Aas et al 2011) There were statistically

significant differences in durability among the feeds

(Table 3) The highest durability value was found for 150 g

gluten kg
1, followed by 0 g gluten kg
1 Overall, except

for 200 g gluten kg
1, the feeds with lower starch content

tended to have lower durability compared with the feeds

with higher starch content

Gluten was not found to have a significant effect on the

WAI The WAI has been correlated with the

microstruc-tural features of the extrudates (Badrie & Mellowes 1991)

In the current work, the WAI values varied between 2.4

and 2.8 (g g
1) Adding 200 g gluten kg
1 increased the

WAI significantly (Table 3), suggesting a structural change

As shown in Table 3, the OAC of ground pellets

increased significantly when 50 g gluten kg
1 was added

and then decreased to a level similar to 0 g gluten kg
1as

the added gluten was further increased to 200 g kg
1 In

general, no significant effect of gluten was observed It waspreviously shown that the OAC of ground extrudates ismostly influenced by protein–oil hydrophobic interactions(Li & Lee 1996) Due to grinding, the effects of the oiluptake rate and pore availability are not included in thismeasurement In contrast, it provided information aboutthe interaction between the oil and the pellet matrix Obvi-ously, the addition of gluten did not change the interactionbetween the matrix and the oil after grinding

SEM micrographs of the outer surfaces and cross-sections

of extruded pellets are shown in Figs 3 and 4, respectively.Figure 3a clearly shows differences in roughness betweenthe samples The product with 200 g gluten kg
1 shows arather smooth surface (Fig 3a) The green areas in Fig 3brepresent the pores, and it is obvious that there are fewerpores on the surface of the pellets with 200 g gluten kg
1compared with the other treatments Image analysisrevealed lower surface porosity for 200 g gluten kg
1(1.8%) compared with the values of 6.9, 8.6 and 6.7% for

Figure 2 Stereomicroscopy images of

cross-sections for feed with 0 (a) and

coated using the pilot-scale coater.

0, 50 and 100 g gluten kg
1, respectively Figure 4 also

shows that the solid matrix is more compartmentalized or

less compact in the sample without gluten

The cryo-SEM micrographs (Fig 5) of the fraction

planes reveal the internal structure of the oil-coated pellet

containing 200 g gluten kg
1 Most of the pores at the

periphery are completely filled with oil, but towards the

core of the pellets, larger voids are not yet filled (Fig 5a)

Unlike the inner part of the pellet (Fig 5c), ridges andfracture points are not visible at the periphery where thecontours are smoother (Fig 5b), which we contribute tothe overlying lipid

The quantitative microstructural data extracted from theXMT analysis are presented in Table 4 No significant dif-ferences in porosity were observed The average air-celldiameter (Dcell) ranged between 319 and 375lm and was

(a)

(b)

.

significantly affected by gluten The lowest Dcell was

observed for pellets without any gluten, and it increased by

15% with 200 g gluten kg
1 Figure 6 shows that most

pore sizes were between 11 and 330lm Porosity smallerthan 11lm could not be detected due to limitations in res-olution The inclusion of gluten led to

Figure 4 SEMs of the cross-section of

(b), showing a more fractured (a) or a

more compact structure (b)

(a)

Figure 5 Cryo-SEM image of a sample

coated using the pilot-scale coater (a)

the image of cross-section is a

compila-tion of a series of representative

representative images of the periphery

and the inner part of the sample,

Table 4 Three-dimensional structural parameters obtained with XMT.

Diet

Porosity (air-cell volume fraction) (volume %)

Average air-cell diameter,

Closed porosity (%)

Connectivity density

vol-ume fraction).

1 less cells smaller than 200lm,

2 more cells in the range from 550 to 700lm, and

3 higher volumetric frequency of structures between 48

and 177lm thick and lower frequency between 200 and

344lm (Fig 7)

When cells expand just after emergence from the die, their

walls stretch and thus become thinner until they rupture

(Trater et al 2005) This explains how an increase in cell size

is accompanied by thinner lamellae between the cells

(Fig 7) Obviously, inclusion of gluten delays rupturing

With the addition of gluten, the surface-to-volume ratio

(SV) tended to decrease, but the differences were not

statis-tically significant (Table 4) The highest SV (0.024lm
1)

was observed for the feed without gluten Even though

there is a (non-significant) trend of slightly higher porosity

with more gluten, smaller cells imply more surface area,

and therefore, it is logical that the samples without gluten

are spread over a greater cell surface area

The DA is a measure of the preferential alignment ofstructures along a particular directional axis in threedimensions (Bellido et al 2006) There is a trend ofdecreasing (P< 0.01) anisotropy with more gluten The

DA values ranged from 1.42 to 1.53, and they were lowerwith the inclusion of gluten, indicating that cells were morespherically shaped (Table 4) Cell anisotropy was more evi-dent at 0 g gluten kg
1

Table 4 shows that all products have an open porousstructure; about 98-99% by volume of all cells is open

This is in the normal range for extruded products maz & Clayton 1992; Bhatnagar & Hanna 1997) Althoughlow in absolute magnitude, the percentages of closed poresincreased (P < 0.05) as the gluten increased from 0 to

(Hicsas-200 g kg
1.Table 4 shows that the connectivity density betweenpores decreased (P< 0.01) with increasing gluten level,although no statistically significant difference was detectedbetween 100 and 200 g gluten kg
1feeds

A series of XMT reconstructed images for the 0, 100 and

200 g gluten kg
1 treatments is given in Fig 8 The darkareas represent the void cells, whereas the continuous solidmatrix, mostly the cell walls, is in orange or white (densermaterial) A change in the pellet microstructure can beobserved with the addition of gluten (Fig 8) The micro-structure of the 0 g gluten kg
1pellet shows elongated cells

at the periphery, caused by axial expansion after emergencefrom the die The centre of the pellet without gluten has anirregular structure

In this work, the microstructure of feed pellets has beenfound to be highly influenced by the replacement of SPC

by vital wheat gluten The changes in microstructure might

be a response to changes in the mechanical behaviour ofthe melt due to changes in the concentration of the twoplant protein sources The overall effect is discussed, andthe origin of the differences is explained based on the struc-ture formation properties of both plant proteins

The increased inclusion of wheat gluten in the pelletsresulted in a marginal increase in protein and a significantincrease in starch The starch level was targeted at

120 g kg
1 DM However, actual measured starch wasbetween 125 and 164 g kg
1for the feed with 50 and 150 g

Figure 6 Cell size distribution in the pellets (average from four

replicates).

Figure 7 Cell wall thickness distribution in the pellets (average

from four replicates).

.

gluten kg
1, respectively (Table 1) The increased starch

level of the feed with 150 g gluten kg
1 might be the

rea-son for the greater expansion of the pellets and therefore

the highest oil infusion (Table 3), despite the changes made

to the process parameters (Table 2) This confirms the

find-ings of Glencross et al (2010), who found that a higher

level of starch results in greater oil uptake due to greater

pellet expansion A similar trend was observed when the

gluten level increased from 0 or 50 to 100 g kg
1(Table 3)

However, this result was not repeated for the 200 g

glu-ten kg
1 inclusion level in the current experiment Based

on the high correlation between the density and maximum

oil infusion, as mentioned earlier, we would have expected

a higher maximum oil infusion at 200 g gluten kg
1 This

indicates hindered oil infusion In the case of 100 and

200 g gluten kg
1, the reduction in sinking is an indicator

that there was no uniform infusion of oil into the product

Besides this effect, the reduced sinking for 150 g

glu-ten kg
1 could also be attributed to the lower density

caused by the high inclusion level of starch compared with

the other feeds

Øverland et al (2009) reported previously that starch

from wheat is a primary component responsible for the

binding properties in extruded diets for salmonids

There-fore, the improved durability of the 150 g gluten kg
1diet(Table 3) in this study could be associated with the higherstarch level in this diet compared with other feeds In gen-eral, the numeric durability values were high for all thefeeds and well within acceptable quality criteria according

to commercial guidelines

The changes in the morphology of the pellets are probablyrelated to the film-forming properties of gluten (Park &Chinnan 1995; Moore et al 2004) Parker et al (1990) andMoss (1974) reported that mixing in bread making causesthe gluten to form a highly extensible network, whichstretches into thin film walls around the growing gas cellsduring leavening The same effect takes place after exitingthe die, during expansion of the matrix, due to flashing off

of the excess steam The higher temperatures used duringextrusion might allow gluten having the same effect as indough development The network formation of hydratedgluten during mechanical deformation was previouslyreported by Bugusu et al (2002) Li & Lee (1996) correlated

(c)

two-dimen-sional horizontal X-ray slice images of

pellets with 0 (a), 100 (b) and 200 g

sections presented are of the whole

pel-let, not the VOI.

the polymerization of gluten through disulphide bond

cross linking with higher WAI values and the formation of

a compact, solid structure Hashimoto et al (2002) reported

higher WAI values with increased gluten concentration in

cassava starch/gluten blends The increase in WAI with the

addition of 200 g gluten kg
1 in our study is therefore an

indicator of network formation by the gluten

The irregular structure of the pellets without gluten

(Fig 8) is probably caused by rupture of smaller cells due

to insufficient strength of the surrounding lamellae Parada

et al (2011) concluded that the solid matrix is more

frag-mented with the addition of fibre by using the average

number of objects per unit length The

compartmentaliza-tion observed in this study (Fig 4) could be caused by the

higher fibre present in the diets with more SPC (i.e less

gluten) The structure around the bubbles at the rim is

solidified quickly, as this part cools off quickest This is in

line with the anisotropy; there was simply not enough time

for relaxation The centre, however, remains mobile for

some time, as it cools more slowly, and if the matrix

mate-rial is not elastic enough to be able to accommodate the

expanding bubbles, it will rupture there, leading to partly

fused smaller cells and large, irregular cells The gluten is

well deformable compared with other types of proteins

with the same level of moisture; thus, the expansion can be

better accommodated Not only does the gluten provide a

highly adaptive environment for the bubbles to grow

dur-ing expansion, but the gluten network remains flexible for

some time, which allows the bubbles to attain a more or

less spherical shape

Rosenquest et al (1975) attributed the small cell size to

a reduction in the strength and stretchability of the dough

When the moisture in the dough flashed to the gaseous

state on extrusion, the walls of the cells broke while the

cells were still small In contrast, as gluten is known to be

an excellent film former, the addition of gluten results in

the lamellae remaining stable over more expansion, thus

leading to larger bubbles and thinner lamellae This

con-firms some of the observations regarding macrostructure

that were made in our previous work (Draganovic et al

2011)

Based on images shown in Fig 2, the greater difference

in maximum oil infusion was expected for the 200 g

glu-ten kg
1 pellets compared with the other treatments

(Table 3) Besides differences in morphology, the rate of oil

infusion is also important here Greater amounts of oil

were used in the infusion tests compared with pilot-scale

coating, which in turn can lead to better oil impregnation

This factor can affect the speed at which the two materials

are intermixed, as well as the distance that one materialcan diffuse into another (Ellis et al 1993) Therefore, for

200 g gluten kg
1, we would expect even lower values formaximum oil infusion in the case of pilot-scale coating(Fig 2)

The results from the current study suggest that oil nation is hindered by the cell structure Although the den-sity would certainly influence these parameters (Barrett &

impreg-Ross 1990), the density is kept constant in this study andtherefore this effect should be attributed to the microstruc-ture The effects of gluten seen here are also found withsome other additions Some natural ingredients can reduceoil uptake because of their film-forming capability (Pedro2009) In the work carried out by Bouchon & Pyle (2004),

it was shown that a more elastic network in a restructuredpotato chip may result in a less permeable outer layer,which is an effective barrier against oil absorption duringfrying

The micrographs used in this study are useful for mating the size and position of the pores on the extrudedsamples (Reitz et al 2008) and allow qualitative observa-tions (Bouchon & Pyle 2004) The outer layer influencesdifferent properties The smooth and regular outer surfaceobtained after extrusion with 200 g gluten kg
1 (Fig 3a)will give a certain resistance to oil impregnation; the oilwill stay on the surface and simply drain off Bouchon &

esti-Pyle (2004) concluded that the evenness of the outer face and the permeability of the outer layer play a funda-mental role in the oil-infiltration pattern

sur-The influence of gluten can be also seen from the resultsfor the diet without gluten Soy protein concentrate isexposed to intolerable levels of stretching during the expan-sion stage, which cannot be accommodated The mechani-cal properties and their timescales do not match thetimescale of the expansion process Thus, a highly irregularand partially ruptured structure results, which allows veryeasy infusion of oil The gluten introduces better deform-ability at processing timescales, and thus, a matrix withgluten can adapt to the expansion, resulting in a structurethat is highly cellular But, slightly more cells are eitherclosed or almost closed (Table 4) These findings are in linewith the results presented by Ghorpade et al (1997) Theyreported that soy protein isolate did not affect the openpore volume of starch-based extrudates; inclusion of glutenabove 200 g kg
1 resulted in a decreased open pore

.

volume Even though the absolute volume of closed cells

for three feeds in this study count for <1.7% of the total

porosity (Table 4), they contribute to the overall density of

the pellets In general, we believe that the potential effect

of closed pores on the reduced oil uptake is minor

com-pared with the effects of surface porosity (Fig 3b) and

other microstructural parameters reported in Table 4 In

accordance with the results from the present study, it has

been shown for tortilla chips that due to capillary pressure,

small narrow pores lead to more fat uptake than wide

pores (Moreira et al 1997) Moreover, the whole pore

pathway has to be considered It has been stated by Saguy

& Pinthus (1995) that long continuous channels lead to

increased fat uptake Both interconnectivity density and

anisotropy give an indication of the pore pathway; less

spherical pores and better interconnectedness between

pores suggest the presence of longer, continuous channels

The values of those two parameters decreased with 200 g

gluten kg
1 (Table 4), and it can be expected that the oil

will penetrate more slowly during oil impregnation, which

will result in more pores (even somewhat open ones) that

are not filled with oil

The observed reduction in fat leakage with 200 g

glu-ten kg
1 (Table 3) could also be attributed to the

differ-ences in skin porosity Once the oil is impregnated in the

pellet with the use of a vacuum, the surface acts as an

effective barrier against its diffusion in the opposite

direc-tion towards the outside Moreover, the leakage of fat

could be also hindered by the lower pore interconnectivity

and less fragmented structure with the 200 g gluten kg
1

pellets For the duration of the test (24 h), there was

lim-ited time for oil diffusion and the differences in fat leakage

would be more pronounced with a longer time

An ideal pellet seems to benefit from inclusion of both,

SPC and gluten Soy protein concentrate provides high

interconnectedness, while gluten provides stronger pellets

When these two components are the main sources of

pro-tein in the diet besides fish meal, their optimization may

lead to an optimal pellet

As stated in Materials and methods, a slight modification

in extruder settings was necessary to obtain pellets with

similar density The 150 g gluten kg
1 sample needed an

adjustment in the barrel temperature (sections 3 and 4),

whereas the 200 g gluten kg
1sample was extruded with a

slightly higher rotation speed (Table 2) The question now

is whether the changes in product properties discussed

above could be attributed to the changes in the extrusionconditions rather than compositional changes

To check for this and the influence of the scale of theprocess, a second twin-screw extruder (TSE 36 HC;Thermo scientific, Staffordshire, UK) with a screw diame-ter of 36 mm and an L/D of 28:1 was used to produce 0and 200 g gluten kg
1 formulations (results not shown).These diets were processed under comparable operatingconditions Here, the inclusion of gluten led to even greaterdifferences in oil uptake using the pilot-scale coater for pel-lets with similar density Moreover, the percentage sinking

of the pellets, fat leakage and the WAI followed exactly thesame trend as with the TX-57 extruder Therefore, we areconfident that the changes in the properties of the finalproduct described in this article are mainly caused by dif-ferences in the technological properties of SPC and gluten

This study shows the relationship between the pellet structure and its techno-functional properties Most of thechanges in techno-functional properties can be explained

micro-by the changes in microstructure Changes in ture were induced by altering the composition through theinclusion of wheat gluten

microstruc-The inclusion of wheat gluten in fish feed pellets leads to

a reduction in oil impregnation and oil uptake during ing, but yields strong, highly porous pellets The pores arestill highly interconnected, albeit slightly less than withoutgluten

coat-Pellets without gluten have an irregular, very open nal structure, while those with gluten have a more regularcellular structure In the radial direction, a larger number

inter-of cells inter-of smaller diameter were observed close to the face of the pellets without gluten, indicating a relativelyporous skin, while gluten yielded a rather smooth, non-por-ous surface

sur-These effects seem to be related to the film-formingproperties of gluten, which are still effectively present atthe high temperatures used during the extrusion process

We gratefully acknowledge the financial support provided

by the Research Council of Norway (project no 197909)for this project and Skretting Aquaculture Research Centrefor provision of facilities The authors would like to thankAdriaan van Aelst and Norbert de Ruijter (WageningenElectron Microscopy Centre, Wageningen University, the

Netherlands) for their assistance with SEM and light

microscopy Kjell Laperre (SkyScan, Antwerpen, Belgium)

is greatly acknowledged for skilful assistance with the

XMT

T (2011) Analysis of pellet degradation of extruded high energy

fish feeds with different physical qualities in a pneumatic feeding

Allan, G.L., Parkinson, S., Booth, M.A., Stone, D.A.J., Rowland,

S.J., Frances, J & Warner-Smith, R (2000) Replacement of fish

meal in diets for Australian silver perch, Bidyanus bidyanus: I.

310.

Anderson, R.A., Conway, H.F & Peplinski, A.J (1970)

Gelatini-zation of corn grits by roll cooking, extrusion cooking and

Badrie, N & Mellowes, W.A (1991) Texture and microstructure

of cassava (Manihot esculenta Crantz) flour extrudate J Food

Barrett, A.H & Ross, E.W (1990) Correlation of extrudate

infus-ibility with bulk properties using image analysis J Food Sci.,

55, 1378–1382.

Bellido, G.G., Scanlon, M.G., Page, J.H & Hallgrimsson, B.

(2006) The bubble size distribution in wheat flour dough Food

Bhatnagar, S & Hanna, M.A (1997) Modification of

Bouchon, P & Pyle, D.L (2004) Studying oil absorption in

Bugusu, B.A., Rajwa, B & Hamaker, B.R (2002) Interaction of

maize zein with wheat gluten in composite dough and bread as

Chanvrier, H., Appelqvist, I.A.M., Bird, A.R., Gilbert, E., Htoon,

A., Li, Z., Lillford, P.J., Lopez-Rubio, A., Morell, M.K &

Top-ping, D.L (2007) Processing of novel elevated amylose wheats:

functional properties and starch digestibility of extruded

Cheng, Z.J.J., Hardy, R.W & Usry, J.L (2003) Plant protein

ingredients with lysine supplementation reduce dietary protein

level in rainbow trout (Oncorhynchus mykiss) diets, and reduce

ammonia nitrogen and soluble phosphorus excretion

Davies, S.J., Morris, P.C & Baker, R.T.M (1997) Partial

substitu-tion of fish meal and full-fat soya bean meal with wheat gluten

and influence of lysine supplementation in diets for rainbow

–328.

Day, L., Augustin, M.A., Batey, I.L & Wrigley, C.W (2006)

Wheat-gluten uses and industry needs Trends Food Sci Tech.,

17, 82–90.

Domenek, S., Brendel, L., Morel, M.-H & Guilbert, S (2004)

Swelling behavior and structural characteristics of wheat gluten

Draganovic, V., van der Goot, A.J., Boom, R & Jonkers, J.

(2011) Assessment of the effects of fish meal, wheat gluten, soy

protein concentrate and feed moisture on extruder system

parameters and the technical quality of fish feed Anim Feed Sci.

Ellis, A., Geselbracht, M., Johnson, B., Lisensky, G & Robinson,

Companion American Chemical Society, Washington, DC.

Feldkamp, I.A., Davis, L.C & Kress, J.W (1984) Practical beam algorithm J Opt Soc of Am.: Optics and Image Science,

Gatlin, D.M., Barrows, F.T., Brown, P et al (2007) Expanding the utilization of sustainable plant products in aquafeeds: a

Ghorpade, V.M., Bhatnagar, S & Hanna, M.A (1997) Structural characteristics of corn starches extruded with soy protein isolate

or wheat gluten Plant Food Hum Nutr (Formerly Qualitas

Gibson, L.G & Ashby, M.F (1997) Cellular Solids-Structure and Properties, 2nd edn Cambridge University Press, Cambridge.

Glencross, B.D., Hawkins, W., Maas, R., Karopoulos, M &

Hauler, R (2010) Evaluation of the influence of different species and cultivars of lupin kernel meal on the extrusion process, pel- let properties and viscosity parameters of salmonid feeds Aqua-

Hashimoto, J.M., Nabeshima, E.H., Cintra, H.S., Guerra Dias,

A.R., Bustos, F.M & Chang, Y.K (2002) Effect of processing conditions on some functional characteristics of extrusion- cooked cassava starch/wheat gluten blends J Sci Food Agric.,

82, 924–930.

Hicsasmaz, Z & Clayton, J.T (1992) Characterization of the pore

132.

Hwang, M.P & Hayakawa, K.I (1980) Bulk densities of cookies

1402.

Content ISO 6865 International Organization for Standardization.

Kaliyan, N & Vance Morey, R (2009) Factors affecting strength and durability of densified biomass products Biomass Bioenerg.,

33, 337–359.

Li, M & Lee, T.C (1996) Effect of cysteine on the functional properties and microstructures of wheat flour extrudates J Ag-

Lin, M.J.Y., Humbert, E.S & Sosulski, F.W (1974) Certain

McCleary, B.V., Gibson, T.S., Solah, V & Mugford, D.C (1994) Total starch measurements in cereal products: interlaboratory evaluation of a rapid enzymatic test procedure Cereal Chem.,

71, 501–505.

Moore, M.M., Schober, T.J., Dockery, P & Arendt, E.K (2004) Textural comparisons of gluten-free and wheat-based doughs,

Moreira, R.G., Sun, X.Z & Chen, Y.H (1997) Factors affecting

Moss, R (1974) Dough microstructure as affected by the addition

of cysteine, potassium bromate, and ascorbic acid Cereal Sci.

Øverland, M., Sørensen, M., Storebakken, T., Penn, M.,

substitut-ing fish meal or soybean meal in diets for Atlantic salmon

digest-ibility, carcass composition, gut health, and physical feed

.

Parada, J., Aguilera, J.M & Brennan, C (2011) Effect of guar

gum content on some physical and nutritional properties of

Park, H.J & Chinnan, M.S (1995) Gas and water vapor barrier

properties of edible films from protein and cellulosic materials.

Parker, M.L., Mills, E.N.C & Morgan, M.R.A (1990) The

poten-tial of immunoprobes for locating storage proteins in wheat

Pedro, B (2009) Chapter 5: Understanding oil absorption during

Science.

Pfeffer, E & Henrichfreise, B (1994) Evaluation of potential

sources of protein in diets for rainbow trout (Oncorhynchus

Reitz, C., Strachan, C & Kleinebudde, P (2008) Solid lipid

extru-dates as sustained-release matrices: the effect of surface structure

Robaina, L., Corraze, G., Aguirre, P., Blanc, D., Melcion, J.P &

Kaushik, S (1999) Digestibility, postprandial ammonia excretion

and selected plasma metabolites in European sea bass

(Dicen-trarchus labrax) fed pelleted or extruded diets with or without

Robin, F., Engmann, J., Pineau, N., Chanvrier, H., Bovet, N & Valle, G.D (2010) Extrusion, structure and mechanical proper-

Rosenquest, A.H., Knipper, A.J & Wood, R.W (1975) Method of producing expanded cereal products of improved texture U.S patent 3927222.

Saguy, I.S & Pinthus, E.J (1995) Oil uptake during deep-fat

_Ibanoglu, Sß (2008) Cauliflower by-products as a new source of dietary fibre, antioxidants and proteins in cereal based ready-to-

Sugiura, S.H., Dong, F.M., Rathbone, C.K & Hardy, R.W (1998) Apparent protein digestibility and mineral availabilities in

Trater, A.M., Alavi, S & Rizvi, S.S.H (2005) Use of non-invasive X-ray microtomography for characterizing microstructure of

Warburton, S.C., Donald, A.M & Smith, A.C (1992) Structure and mechanical properties of brittle starch foams J Mater Sci.,

27, 1469–1474.

1,2 2 2 2,3 2 2 2

2 1

Animal Science College, South China Agricultural University, Guangzhou, China; 2 School of Life Science, Sun Yat-Sen

University, Guangzhou, China;3 Key Laboratory of Freshwater Fisheries and Germplasm Resources Utilization, Ministry

of Agriculture, Freshwater Fisheries Research Center, Chinese Academy of Fishery Sciences, Wuxi, China

In order to investigate the effects of lysine and dissolved

oxygen on grass carp, the grass carp were fed with 13, 15

and 17 g kg
1 lysine diet at about 6 mg L
1 (high

dis-solved oxygen, HO group) or 3.5 mg L
1 (low dissolved

oxygen, LO group) dissolved oxygen level, for 8 weeks

The fish were fed to apparent satiation by hand The

results showed that apparent digestibility of protein, energy

and dry matter were decreased significantly when grass

carp were fed at 3.5 mg L
1 dissolved oxygen, and feed

intake (FI) was also inhibited by low dissolved oxygen

(P < 0.05) Weight gain, protein retention, protein

effi-ciency, feed conversion ratio and amino acid retention of

fish at 6 mg L
1 dissolved oxygen level were significantly

improved at 3.5 mg L
1 dissolved oxygen level (P< 0.05)

Weight gain, protein and amino acid retention, and feed

efficiency of grass carp at the two dissolved oxygen levels

were significantly improved by lysine supplementation

(P < 0.05) The dietary lysine level and dissolved oxygen of

water had an interaction effect on feed conversion ratios

(P < 0.05) Grass carp fed at low dissolved oxygen level

showed lower liver protein and fat contents Plasma

aspar-tate aminotransferase (AST) activity of grass carp fed at

3.5 mg L
1 dissolved oxygen level was significantly

increased compared to 6 mg L
1 dissolved oxygen level

(P < 0.05) Our results show that low dissolved oxygen

level of water is harmful to the liver of grass carp

KEY WORDS: Ctenopharyngodon idella, lysine, oxygen

Received 28 December 2011; accepted 14 November 2012

Correspondence: Yong-Jian Liu, Institute of Aquatic Economic Animals, School of Life Science, Sun Yat-sen University, 135 Xin’gang Xi Road, Guangzhou 510275, China E-mails: edls@mail.sysu.edu.cn; ganlian@

scau.edu.cn

Grass carp (Ctenopharyngodon idella) has a long history

in aquaculture and is one of the most important speciescultured in inland water bodies in China After silvercarp, grass carp currently has the largest production infreshwater aquaculture globally It constitutes 7.18% ofthe world aquaculture production (FAO 2010) Low dis-solved oxygen is a type of stress frequently found in grasscarp farms characterized by high fish densities and pol-luted fresh waters in China Dissolved oxygen (DO) is themost important factor controlling growth, and a long con-stant DO concentration below a critical level is considered

to depress FI, growth and food conversion efficiency(Thetmeyer et al 1999; Buentello et al 2000; Pichavant

et al 2000; Foss et al 2003; Tran-Duy et al 2008) Fishtry to maintain oxygen delivery in the face of reductions

in water oxygen levels If oxygen delivery is compromisedand tissue oxygen levels fall, then energy expenditure isreduced and anaerobic metabolism is up-regulated Sometabolism was changed when fish was exposed to acuteenvironmental hypoxia (Dunn & Hochachka 1986; Bouti-lier et al 1988; Gracey et al 2001, 2011; Delaney & Kle-sius 2004)

Lysine is the most limiting amino acid in plant proteinmeal such as canola meal and cotton meal, which areimportant feed stuffs locally available for formulating grasscarp diets Lysine levels can no longer satisfy the require-ments in the commercial diets because plant protein meal

.

doi: 10.1111/anu.12030 .2013 19; 860–869

Aquaculture Nutrition

was used as a major protein resource for grass carp (Yang

et al 2010) When lysine was deficient, the growth

perfor-mance of grass carp was depressed, and the feed conversion

efficiency was very poor (Wang et al 2005)

Although dietary lysine and dissolved oxygen of water

are very important for cultured species, most grass carps

were fed with lysine-deficient diets in low dissolved oxygen

ponds in commercial farms of China However, the

rela-tionship between fish nutrition and environmental factors,

especially between dietary lysine and dissolved oxygen of

water, is still less studied So our objective was to determine

whether there is any interaction effect between dissolved

oxygen and dietary lysine levels in grass carp

Three diets were prepared (Table 1) The basal diet

con-tained the minimum level of lysine, 13 g kg
1 DM, from

soybean meal, canola meal, cotton meal, rice bran meal

and wheat flour (Zhuhai Shihai Feed Corporation Ltd,

Zhuhai, China) Lysine concentration was increased in two

steps of approximately 2 g lysine kg
1 DM each, with

lysine originating from L-lysine HCL (78% L-lysine; CJ

Co., Ltd, Liaocheng, China) The maximum intended lysine

concentration was 17 g kg
1 DM, which is optimally

required by grass carp Lysine supplementation was carried

out by substituting L-glutamic acid for L-lysine HCL

Methionine was supplied to meet the requirement along

with MHA.Ca (84%, Novus International Inc., Zhibo,

China) Amino acid composition of experimental diets is

shown in Table 2 All diets were isoenergetic

All dry ingredients were finely ground, weighed, mixed

manually for 5 min and then transferred to a Hobart mixer

(A-200T Mixer Bench Model unit, Resell Food Equipment

Ltd, Ottawa,ON, Canada) for another 15-min mixing Soya

lecithin was added to a preweighed premix of soy oil and

mixed until homogenous The oil mix was then added to

the Hobart mixer slowly while mixing was still continuing

All ingredients were mixed for another 10 min Then,

dis-tilled water (about 300 g kg
1diet) was added to the

mix-ture to form dough The wet dough was placed in a

pelletizer (Institute of Chemical Engineering, South China

University of Technology, Guangzhou, China) and pelleted

through a 1.25-mm die The diets were dried with forced

air at 20°C for 24 h, and the moisture was reduced to

about 100 g kg
1 The dry pellets were placed in plastic

bags and stored in a deep freezer at
20 °C until used

Grass carp juvenile from our facilities were used in this iment, and their initial wet weights were 4.11 0.03 g Beforethe experiment, the fish were acclimated to the experimentalconditions for 2 weeks and fed a commercial diet containing

exper-300 g kg
1protein and 40 g kg
1lipid to satiation five healthy fish were randomly distributed to each of 18experimental fibreglass tanks (98 L9 48 W 9 42 H cm,water volume of 200 L) connected to a recirculation system

Twenty-Table 1 Formulation and approximate composition of practical diets for grass carp

285; CaHPO4.2H2O, 250; FeSO4.7H2O, 200; MnSO4.H2O, 25; CoSO4.7H2O, 25; CaIO3, 25; CuSO4.5H2O, 15; Na2SeO3, 10 (Gu- angzhou Chengyi Aquatic Technology Ltd, Guangzhou, China).

A, 1500 IU; vitamin E, 40; vitamin D3, 2000 IU; menadione, 6; idoxine, 4; cyanocobalamin, 2; biotin, 2; calcium pantothenate, 25; folic acid, 2; niacin, 12; inositol, 50(Guangzhou Chengyi Aqua- tic Technology Ltd, Guangzhou, China).

Gu-angzhou, China).

The study was carried out at two levels of dissolved

oxy-gen Dissolved oxygen treatments were a high DO and a

low DO Dissolved oxygen levels inside the tanks were

established by means of aeration and water flow rates

through the tanks Each tank assigned to the high DO level

was aerated with one air stone Tanks assigned to the low

DO level were not aerated After the fish were allocated to

the tanks, water flow rates were set at 2 L min
1 Then, the

flow rates in the tanks assigned to the low DO level were

gradually reduced to 0.22 L min
1within the next 5 days

Dissolved oxygen concentrations and temperature at the

inlet, inside (central point) and outlet of each tank were

measured six times per day, at around 7:00, 9:00, 11:00,

13:00, 15:00 and 17:00 using oxygen meters (YSI500; USA)

The dissolved oxygen concentration during the feeding

per-iod is shown in Fig 1 On 26 May, 2 June, 9 June, 16 June,

23 June, 30 June, 7 July and 14 July, DO concentrations

were also measured every 2 h from 17:30 to 17:30 the next

day to assess the diurnal variation in DO concentration (See

Fig 2) Water samples were collected for total ammonia

nitrogen analyses every 2 days The dead fish also was

col-lected and weighed to correct final total fish body weight

The water was oxygenated and passed through artificial

sponge (3 cm thickness), coral sand (25 cm thickness) and

active carbon filter (25 cm thickness) to remove chlorine

During the trial period, the diurnal cycle was 12-h

light/12-h dark Water quality parameters monitored weekly were

as follows: temperature, 29.1 2.4°C; pH, 7.9  0.09,respectively

The fish were fed manually thrice per day to apparentsatiation for 8 weeks Faeces were collected daily duringthe last 2 weeks as described by Wang et al (2005) Faecestank
1 was dried at 105°C and stored at
70 °C for thedetermination of digestibility with Y2O3as indicator

At the beginning of the feeding trial, 18 fish were randomlysampled from the initial fish and killed for analyses ofwhole-body composition At the end of the 56-day experi-ment, 10 fish from each tank were randomly collected forproximate analysis, 4 for the analysis of whole-body com-position and 6 were anaesthetized with tricaine methanesulphonate (MS222) (50 mg L
1) for blood collection and

to obtain weights of individual whole body, viscera, liverand intraperitoneal fat White muscle from both sides ofthe fillets without skin and liver were dissected and frozenimmediately in liquid nitrogen and stored at
70 °C until

Table 2 Amino acid composition of experimental diets for grass

Diet

Essential amino acids

Tryptophan is not detected.

As an analog of methionine, MHA.Ca cannot be detected by the

amino acid analyser, so methionine value was analysed as the

sum of MHA.Ca and methionine.

1 2 3 4 5 6 7

26-May 2-Jun 9-Jun 16-Jun 23-Jun 30-Jun 7-Jul 14-Jul

–1 )

Date

LO HO

Figure 1 Variations in dissolved oxygen (DO) level throughout the

0 1 2 3 4 5 6 7

Figure 2 Diurnal variations in dissolved oxygen (DO) level.

.

used The plasma was separated by centrifugation and also

stored at
70 °C until analysed

Diets and fish samples (including white muscle and liver)

were analysed in triplicate for proximate composition

Crude protein, crude lipid, moisture, crude ash and gross

energy were determined following standard methods

(AOAC 1984) Crude protein (N9 6.25) was determined

by the Kjeldahl method after acid digestion using an Auto

Kjeldahl System (1030-Auto-analyzer; Soxtec System, Tecator

AB, Sweden) Crude lipid was determined by the ether

extraction method using a Soxtec System HT (Soxtec System

HT6, Tecator) Moisture was determined by oven drying at

105°C for 24 h Crude ash was determined by incineration

in a muffle furnace at 550°C for 24 h Gross energy was

determined using an adiabatic bomb calorimeter Amino

acids were analysed following acid hydrolysis using

high-pressure liquid chromatography (HPLC; Hewlett Packard

1090, Palo Alto, CA, USA) The concentrations of dietary

and faecal Y2O3 were determined by inductively coupled

plasma atomic emission spectrophotometer [ICP; model:

IRIS Advantage (HR), Thermo Jarrel Ash Corporation,

Boston, MA, USA) after perchloric acid digestion (Bolin

et al 1952) The concentrations of total plasma protein

(TP), albumin (ALB), cholesterol (CHO), triacylglycerol

(TG), glucose (GLU), aspartate aminotransferase (AST),

alanine aminotransferase (ALT), high-density lipoprotein

(HDL), low-density lipoprotein (LDL), urea-N and

gluta-mate dehydrogenase (GLDH) were determined using an

automatic blood analyser (Hitachi 7170A, Hitachi 7170A,

Hitachi Ltd, Japan) from a clinical laboratory

All data are presented as means SEM TheSPSSsoftware

version 13.0 for Windows of GLM procedure (SPSS Inc.,

Chicago, IL, USAVer13.0, USA) was used to conduct

fac-torialANOVAto determine the effects of dietary protein

con-tent, crystal amino acid supplementation and interaction

between the two factors When interaction between protein

level and amino acid supplementation was statistically

sig-nificant for a particular response, differences among

pro-tein levels within each diet type were determined using

Tukey’s mean separation Treatment effects and

interac-tions were considered significant at P< 0.05

Mean total ammonia nitrogen concentrations under HO

and LO groups were 0.61 0.15 mg L
1 and

0.59 0.11 mg L
1 respectively Total ammonia nitrogenconcentrations in high oxygen tanks were higher than inlow oxygen tanks, but there was no difference Fish readilyaccepted the experimental diets, and survival rate was veryhigh during the 56-day feeding trial There were no signifi-cant differences in survival among fish fed all the diets(Table 3) After 8 weeks’ feeding trial, the final bodyweight, weight gain (WG), specific growth rate (SGR),nitrogen retention (NR), lipid retention (LR) and proteinefficiency ratio (PER) of grass carp were significantlydepressed by the low dissolved oxygen (P< 0.001) and alsowere significantly improved with lysine supplementation(P< 0.05) except lipid retention Feed intake was signifi-cantly increased with an increase in dissolved oxygen(P< 0.001), and not affected by dietary lysine content.Feed conversion ratio (FCR) also was improved with anincrease in lysine and dissolved oxygen (P< 0.05), butinteraction was found between dissolved oxygen and dietlysine level (P< 0.05)

The proximate compositions of whole body, white cle and liver of the grass carp are shown in Table 4 Noeffects of dissolved oxygen on whole-body crude protein,moisture, lipid and ash contents of the fish were found.The whole-body lipid content was significantly decreasedwhen fish was fed diets with lysine supplementation(P< 0.001) The liver protein and lipid contents were sig-nificantly reduced when fish were fed at the low dissolvedoxygen (P< 0.05) The composition of white muscle wasnot affected by dissolved oxygen and lysine level

mus-Condition factor, hepatopancreasomatic index (HSI),intraperitoneal fat (IPF) and viscerosomatic index (VSI) ofgrass carp fed experimental diets are presented in Table 4.VSI and IPF decreased with increasing dietary lysine levelsamong diet treatments VSI, IPF and HSI of grass carp atthe low dissolved oxygen were significantly higher thanthose at high dissolved oxygen (P< 0.05)

Apparent digestibilities of dry matter, protein and energyprovided in Table 5 were significantly affected by dissolvedoxygen concentration (P< 0.001), but not dietary lysinelevel The results indicated that apparent digestibilities ofdry matter, protein and energy were significantly reduced

at the low dissolved oxygen concentration No interactionwas detectable between the two experimental factors withregard to digestibility

Plasma biochemical parameters are provided in Table 6

No effects were found with increasing dietary lysine level.The results indicated that AST, ALT, TG, GLU, urea andGLDH of grass carp fed at low dissolved oxygen were sig-nificantly higher than those of grass carp fed at high dis-

solved oxygen (P< 0.001) High-density lipoprotein and

LDL were significantly decreased when grass carp were fed

at low dissolved oxygen (P< 0.05) No interactions were

found

Amino acid retentions are provided in Table 7 The

results indicated that amino acid retention except

methio-nine were not significantly affected by dietary lysine level,

but it showed an increasing trend with increasing lysine

level Threonine, proline, glycine, alanine, cystine, valine,

methionine, isoleucine, leucine, phenylalanine and histidine

retention of grass carp at low dissolved oxygen were

signifi-cantly lower than those at high dissolved oxygen

(P < 0.05) No interaction was found

When grass carp was fed at low dissolved oxygen, apparent

digestibility of grass carp was significantly reduced

Axels-son & Fritsche (1991) showed that acute hypoxia imposed

an increased visceral vascular resistance in the blood flow

in Gadus morhua, leading to a reduction in the coeliac and

mesenteric artery blood supply Similar results were found

in Dicentrarchus labrax (Axelsson et al 2002) It may be

expected that a reduction in blood flow can depress the

digestion of fish Tran-Duy et al (2011) had demonstrated

that apparent digestibility of Oreochromis niloticus fed at

low dissolved oxygen was reduced

Decreased oxygen availability is also considered a major

factor in determining food intake Grass carp showed a

reduced appetite and growth performance when fed at low

dissolved oxygen Similar results have been obtained from

Oncorhynchus mykiss (Pedersen 1987; Glencross 2009),

D labrax L (Thetmeyer et al 1999), Scophthalmus

maxi-mus (Pichavant et al 2000), Anarhichas minor Olafsen

(Foss et al 2003), Ictalurus punctatus (Buentello et al

2000), O niloticus (Tran-Duy et al 2008), Morone

saxatilis (Brandt et al 2009), Hippoglossus hippoglossus

L (Thorarensen et al 2010); all these fish experiencedreduced growth It is not surprising that fish show poorappetite during long hypoxia When dissolved oxygen avail-ability of water drops to a level that cannot support aero-bic metabolism, fish will shift to anaerobic pathways forenergy production Subsequently, metabolic depressionoccurs to minimize energy expenditure Fish reduce or stopfeeding completely during hypoxic conditions, presumablybecause food digestion is energetically demanding Acquisi-tion of food and its digestion and assimilation are majorenergy expenditures (up to 60%) of fishes (Van Dam &

Pauly 1995)

Feed efficiency was affected by the dissolved oxygen, andfish always showed good feed efficiency when fed at enoughdissolved oxygen water (Bergheim et al 2006) In ourstudy, FCR of grass carp was significantly reduced whendissolved oxygen was increased, and the lipid, proteinretention, protein efficiency ratio and amino acid retentionalso were significantly improved Blood urea and GLDHactivity were major indicators of amino acid metabolism(Kim et al 1987; Regnault 1987; Encarnacß~ao et al 2004)

The blood urea level and the activity of GLDH were icantly higher when grass carp was fed at low dissolvedoxygen level It indicated that protein synthesis of grasscarp was inhibited by low dissolved oxygen Smith et al

signif-(1996) found that when crucian carp were exposed to 48-hanoxia, there was more than a 56% reduction in proteinsynthesis rate in liver, 52% in red muscle and 56% in whitemuscle

Although there was a controversy about utilization ofcrystal amino acids for fish, the growth performance andamino acid retention of grass carp fed at both dissolvedoxygen levels were improved with lysine HCL supplementa-tion in our study The same result was obtained by Yang

et al.(2010) Some researchers also have demonstrated thatLabeo rohita (Mukhopadhayay & Ray 1999; Sardar et al

2009), O niloticus (Furuya et al 2004), O mykiss (Cheng

Table 5 Apparent digestibility of nutrients and energy of juvenile grass carp fed with lysine supplementation at different oxygen levels

et al 2003), Pagrus major (Takagi et al 2002) and tus mesopotamicus (Abimorad et al 2009) fed diets supple-mented with crystal lysine and methionine had bettergrowth performance In our study, FCR of grass carp alsowas significantly reduced with increasing lysine, and therewas an interaction effect between lysine level and dissolvedoxygen It indicated that high dissolved oxygen and bal-anced amino acid diet were helpful to obtain low FCR andgood growth performance of fish.

Piarac-Hepatopancreasomatic index of grass carp fed at low solved oxygen was significantly higher than that of those fed

dis-at high dissolved oxygen, but the protein and lipid contents

of liver showed an opposite trend It indicated that the liver

of grass carp was impaired when fed at low dissolved gen Aspartate aminotransferase and alanine aminotransfer-ase are usually used as general indicators of the functioning

oxy-of vertebrate liver (Wroblewski & Ladue 1956) High ASTand ALT generally indicate the damage or weakening ofnormal liver function ALT and AST were often used asmarkers of hepatocellular injury (Seymen et al 1999; O’Bri-

en et al 2000; Pan et al 2010) And in our study, plasmaALT and AST of grass carp fed at low dissolved oxygenwere significantly increased, and it demonstrated that liverwas impaired when grass carp was cultured at low dissolvedoxygen The same conclusion was obtained in Hyphessobry-con callistusBoulenger (Pan et al 2010)

Blood glucose level was significantly higher when grasscarp were fed at low dissolved oxygen level Acute hypoxiawas known to increase the levels of catecholamines, activat-ing glycogenolysis and gluconeogenesis with a net result ofincreasing plasma glucose levels (Wright et al 1989) Simi-lar results were found on Sparus aurata (Henrique et al.1998), Nephrops norvegicus (Schmitt & Uglow 1998), O nil-oticus(Delaney & Klesius 2004) and I punctatus (Thomas

L et al 2007) The glucose concentration increment wasrelated to the mobilization of energy storage under stressfulconditions of low oxygen availability, as a source of fuelfor anaerobic metabolism During exposure to acutehypoxia, cardiac ATP concentration of Tilapia (Oreochr-omis hybrid sp.) was unchanged compared with normoxiaand anaerobic glycolysis contributed to ATP supply as evi-denced by considerable accumulation of lactate in the heartand plasma (Speers-Roesch et al 2010) When Gillichthysmirabilis was exposed to short hypoxia, Gracey et al.(2001) showed that genes involved in the glycolytic meta-bolic pathway, muscle contraction and locomotion of

G mirabilis are all down-regulated in the muscle cells Onthe other hand, several genes involved in gluconeogenesiswere up-regulated in the liver during hypoxia

The results of the present study showed that FI and growth

of grass carp were depressed when fed at low dissolved

oxygen, and the liver may be impaired The growth

perfor-mance of grass carp was improved when it was fed a diet

with lysine supplementation

Abimorad, E.G., Favero, G.C., Castellani, D., Garcia, F & Carneiro,

D.J (2009) Dietary supplementation of lysine and/or methionine

on performance, nitrogen retention and excretion in pacu Piaractus

AOAC (1984) Official Methods of Analysis pp 1141 Association

of Official Analytical Chemists, Arlington, VA.

Axelsson, M & Fritsche, R (1991) Effects of exercise, hypoxia

and feeding on the gastrointestinal blood flow in the Atlantic

Axelsson, M., Altimiras, J & Claireaux, G (2002) Post-prandial

blood flow to the gastrointestinal tract is not compromised

dur-ing hypoxia in the sea bass Dicentrarchus labrax J Exp Biol.,

205, 2891–2896.

Bergheim, A., Gausen, M., Næss, A., Hølland, P.M., Krogedal, P.

& Crampton, V (2006) A newly developed oxygen injection

Bolin, D.W., King, R.P & Klosterman, E.W (1952) A simplified

method for the determination of chromic oxide when used as an

Boutilier, R., Dobson, G., Hoeger, U & Randall, D (1988) Acute

exposure to graded levels of hypoxia in rainbow trout (Salmo

gairdneri): metabolic and respiratory adaptations Respir

Cheng, Z.J.J., Hardy, R.W & Usry, J.L (2003) Plant protein ingredients with lysine supplementation reduce dietary protein level in rainbow trout (Oncorhynchus mykiss) diets, and reduce ammonia nitrogen and soluble phosphorus excretion Aquacul-

Delaney, M.A & Klesius, P.H (2004) Hypoxic conditions induce Hsp70 production in blood, brain and head kidney of juvenile

Dunn, J & Hochachka, P (1986) Metabolic responses of trout (Salmo gairdneri) to acute environmental hypoxia J Exp Biol.,

123, 229–242.

Bureau, W & Bureau, D.P (2004) Diet digestible energy content affects lysine utilization, but not dietary lysine requirements of rainbow trout (Oncorhynchus mykiss) for maximum growth.

Furuya, W.M., Pezzato, L.E., Barros, M.M., Pezzato, A.C., uya, V.R.B & Miranda, E.C (2004) Use of ideal protein con- cept for precision formulation of amino acid levels in fish-meal- free diets for juvenile Nile tilapia (Oreochromis niloticus L.).

.

Glencross, B (2009) Reduced water oxygen levels affect maximal

feed intake, but not protein or energy utilization efficiency of

Gracey, A.Y., Troll, J.V & Somero, G.N (2001) Hypoxia-induced

gene expression profiling in the euryoxic fish Gillichthys mirabilis.

Gracey, A.Y., Lee, T.H., Higashi, R.M & Fan, T (2011)

Hypoxia-induced mobilization of stored triglycerides in the

Henrique, M., Gomes, E., Gouillou-Coustans, M., Oliva-Teles, A.

& Davies, S (1998) Influence of supplementation of practical

diets with vitamin C on growth and response to hypoxic stress

Kim, K., Kayes, T & Amundson, C (1987) Effects of dietary

tryptophan levels on growth, feed/gain, carcass composition and

liver glutamate dehydrogenase activity in rainbow trout (Salmo

Mukhopadhayay, N.R & Ray, A.K (1999) Improvement of

qual-ity of sal (Shorea robusta) seed meal protein with supplemental

amino acids in feed for rohu, Labeo rohita (Hamilton),

O’Brien, P., Slaughter, M., Swain, A., Birmingham, J., Greenhill,

R., Elcock, F & Bugelski, P (2000) Repeated acetaminophen

dosing in rats: adaptation of hepatic antioxidant system Hum.

Pan, C.H., Chien, Y.H & Wang, Y.J (2010) The antioxidant

capacity response to hypoxia stress during transportation of

characins (Hyphessobrycon callistus Boulenger) fed diets

Pedersen, C.L (1987) Energy budgets for juvenile rainbow trout at

Pichavant, K., Person-Le-Ruyet, J., Le Bayon, N & Severe, A.

(2000) Effects of hypoxia on growth and metabolism of juvenile

Regnault, M (1987) Nitrogen excretion in marine and fresh-water

Sardar, P., Abid, M., Randhawa, H.S & Prabhakar, S.K (2009)

Effect of dietary lysine and methionine supplementation on

growth, nutrient utilization, carcass compositions and

haemato-biochemical status in Indian Major Carp, Rohu (Labeo rohita

Schmitt, A.S.C & Uglow, R.F (1998) Metabolic responses of

Seymen, H.O., Seven, A., Civelek, S., Yigit, G., Hatemi, H &

Bur-cak, G (1999) Evaluation of antioxidant status in liver tissues:

effect of iron supplementation in experimental hyperthyroidism.

Smith, R.W., Houlihan D.F., Nilsson, G.E & Brechin, J.G (1996) Tissue-specific changes in protein synthesis rates in vivo

Speers-Roesch, B., Sandblom, E., Lau, G.Y., Farrell, A.P & ards, J.G (2010) Effects of environmental hypoxia on cardiac energy metabolism and performance in tilapia Am J Physiol.

Takagi, S., Shimeno, S., Hosokawa, H & Ukawa, M (2002) Effect of lysine and methionine supplementation to a soy protein

Thorarensen, H., Gustavsson, A.o., Mallya, Y & Gunnarsson, S (2010) The effect of oxygen saturation on the growth and feed conversion of Atlantic halibut (Hippoglossus hippoglossus L.).

Tran-Duy, A., Schrama, J.W., van Dam, A.A & Verreth, J.A.J (2008) Effects of oxygen concentration and body weight on max- imum feed intake, growth and hematological parameters of Nile

Tran-Duy, A., van Dam, A.A et al (2011) Feed intake, growth and metabolism of Nile tilapia (Oreochromis niloticus) in rela-

–744.

Van Dam, A.A & Pauly, D & Schrama, J.W (1995) Simulation

of the effects of oxygen on food consumption and growth of Nile

Wang, S., Liu, Y., Tian, L., Xie, M., Yang, H., Wang, Y & Liang, G (2005) Quantitative dietary lysine requirement of

Wright, P., Perry, S & Moon, T (1989) Regulation of hepatic coneogenesis and glycogenolysis by catecholamines in rainbow

transaminase(SGP-T) in hepatic disease: a preliminary report.

Yang, H., Liu, Y., Tian, L., Liang, G & Lin, H (2010) Effects of supplemental lysine and methionine on growth performance and body composition for grass carp (Ctenopharyngodon idella) Am.

1 2 3

1

Oceanography Department, Faculty of Science, Alexandria University, Alexandria, Egypt; 2 Zoology Department, Faculty

of Sciences, Damietta University, Damietta, Egypt; 3 National Institute of Oceanography and Fisheries, Alexandria,

Egypt

Two experiments were conducted to investigate the effects of

feed colour on the performance of Nile tilapia (Oreochromis

niloticus) larvae and fingerlings In the first experiment,

trip-licate groups of newly hatched larvae (0.01 g fish
1) were

stocked in 40 L glass aquaria at a density of 2 fish L
1 The

fish were fed a test diet (400 g kg
1crude protein) with six

different colours (dark blue, dark green, red, dark brown,

yellow and light brown) for 60 days The best performance

and survival were achieved in fish fed on dark-coloured diets,

while light-coloured diets (yellow and light brown) resulted

in inferior performance Dark diets also produced higher

body protein than light diets Body water, lipids and ash

showed irregular trends In the second experiment, triplicate

groups of Nile tilapia fingerlings (5.30 g fish
1) were stocked

in 140-L aquaria, in a recirculating indoor system The fish

were fed a test diet (350 g kg
1crude protein) with the same

colours used in the larval trial, for 60 days Growth rates,

feed efficiency, survival and body composition were not

sig-nificantly affected by feed colours These results suggest that

Nile tilapia larvae are visual feeders, and they prefer

dark-coloured diets to light-dark-coloured diets, while fingerling fish

showed no preference to diet colours

KEY WORDS: feed colour, fingerlings, larvae, Nile tilapia

Received 21 September 2012; accepted 26 November 2012

Correspondence: A.-F.M El-Sayed, Oceanography Department, Faculty

of Science, Alexandria University, Alexandria, Egypt.

E-mail: afmelsayed@gmail.com

During the past two decades, tilapia culture has been

shar-ply growing throughout the world, particularly in Asia,

Africa and the Americas As a result, the global production

of farmed tilapia has increased from 383 654 tonnes in

1990 representing 4.5% of total farmed fish production to

3 497 391 tonnes in 2010, representing 8.9% of totalfarmed fish production, with an average annual growth of13.5% (FAO 2012) This trend has created a gap betweenseed supply and farmer’s demand

The global, rapid industrialization of tilapia production

in recent years has also led to gradual shift in tilapia ture from extensive and semi-intensive systems to moreintensive farming practices, with an increasing dependence

cul-on formulated feeds (El-Sayed 2007) Therefore, the majorchallenge facing tilapia culture industry is the production

of sufficient amounts of quality seeds and formulation ofappropriate, cost-effective feeds Proper feed and feedingmanagement is a necessary tool for successful tilapia cul-ture practices In another word, the profitability of a tila-pia farm is directly related to the amount of feedconsumed by the fish Highest feed consumption and feedutilization efficiency generally leads to highest growthrates, and, in turn, maximum profitability

Therefore, feeding stimulants could play an importantrole in improving feed consumption and growth rates

However, feeding and dietary preferences of tilapiasdepend on tilapia species and size (Turker et al 2003),time of day (Fortes-Silva et al 2011), photoperiod (Bi-swas & Takeuchi 2002; El-Sayed & Kawanna 2004; Rad

et al 2006), diet form and size (Santiago et al 1987),diet colour (Arumugam 1997; El-Sayed 2004), back-ground colour (Volpato & Barreto 2001; Volpato et al

2004; Luchiari et al 2007) and culture conditions(El-Sayed 2006) The effects of most of these factors ontilapia culture, feeding and behaviour have been wellstudied

The feeding success of visual feeders depends mainly

on the contrast between the prey and its background

.

doi: 10.1111/anu.12031 .2013 19; 870–876

Aquaculture Nutrition

(Fiksen et al 1998; Utne-Palm 1999), which enables

bet-ter prey visualization and subsequently improved feed

intake (Martin-Robichaud & Peterson 1998) The

response of these fish to background colour is also

spe-cies specific and may vary according to life stage of the

same species For example, Van der Salm et al (2005)

found that Mozambique tilapia Oreochromis mossambicus

can change their body colour in response to a change in

background colour The effects of background colour on

chromatophore-regulating hormones in tilapia have also

been reported (Gro¨neveld et al 1995) Furthermore,

background colour has been shown to affect the

behav-iour (aggressiveness, swimming, respiratory frequencies,

stress and nest building) of O niloticus (Fanta 1995;

Volpato & Barreto 2001)

However, the effects of feed colour on growth rates and

feeding efficiency of farmed tilapia have not yet been well

investigated Very few studies have considered the effects of

feed colour on the performance of different tilapia species,

with contradictory results For example, Arumugam (1997)

reported that tilapia hybrid fry fed encapsulated feeds with

different colours (red, orange and green) showed no

prefer-ence to diet colours On the other hand, Jegede & Olusola

(2010) reported that Tilapia zillii fed light-coloured diets

(yellow and light-green) had better growth and feed

effi-ciency than those fed on dark-coloured diets

On the contrary, some other authors reported that

dark-coloured diets provided better growth and feed efficiency of

juvenile Nile tilapia than lighter diets (El-Sayed 2004;

Jeg-ede & Olusola 2010) These discrepancies may have been

attributed to the difference in fish species, sizes and life

stages, background colour of culture units, feed

composi-tion and culture condicomposi-tions Therefore, more, long-term

studies should be conducted to verify the effects of feed

colours on the performance of farmed tilapia, especially

Nile tilapia, at different life stages, and under different

cul-ture systems and environmental settings

This study was carried out, in two consecutive

experi-ments, to investigate the effects of feed colour on growth,

feed utilization efficiency and survival of Nile tilapia

(Ore-ochromis niloticus L.) fry (experiment 1) and fingerlings

(experiment 2)

Newly hatched, mixed-sex Nile tilapia (Oreochromis

niloti-cus) larvae (average initial weight of 0.01 g) used in the

first experiment were obtained from a commercial tilapiahatchery at Farascore, Damietta, Egypt The larvae weredistributed into 40 L rearing glass aquaria, filled with de-chlorinated tap water, at a density of 80 fish per aquarium(2 fish L
1) Each aquarium was equipped with a filter con-nected to an air stone for aeration The aquaria werecleaned each morning, and about 25% of the culture waterwas replaced with fresh, dechlorinated water of similartemperature

Nile tilapia fingerlings used in the second experimentwere obtained from a commercial hatchery at Edko, Beha-ira Governorate, Egypt Triplicate groups of fish (averageinitial weight of 5.30 g) were stocked in glass aquaria(140 L each) connected together in a closed, recirculatingsystem, containing a biological filter, at a density of 20 fishper aquarium

Fish in each experiment were acclimatized to the mental conditions for 1 week, during which they were fedthe test diets Lighting in the culture facilities was set at12:12 light/dark cycle, using fluorescent lamps Water qual-ity parameters including, dissolved oxygen (DO), Ammonia(NH4-N), Nitrates (NO3-N), Nitrites (NO2-N) and pH weremonitored weekly, using HACH test kit (Loveland, CO,USA) The average values of these parameters are summa-rized in Table 1

experi-A 400 g kg
1 crude protein (cp) basal test diet (Table 2)was prepared by weighing and mixing the dry ingredients.Then, the fish oil and soybean oil were mixed togetherand added to the dry mix drop by drop with continuousmixing The mix was then divided into six parts, each ofthem was coloured with one of six different colours (darkblue, dark green, dark brown, red and yellow), whereasthe 6th part (light brown) was kept non-coloured and

Table 1 Water quality measurements (mean  SE) of Experiments

served as a control diet Diets were coloured using

com-mercial human food colourants (Ajanta Food Colours,

Gurgaon, Haryana, India) These colourants are widely

used for colouration of human food, drinks, syrup and

dessert Each food colourant was dissolved in a

250 mL kg
1feed) The water was added gradually with

mixing, until the diets were moulded into small balls The

diets were then passed through a meat grinder to form

spaghetti-like threads, spread on aluminium foil plates and

dried in an oven at 60°C for 24 h Finally, the dried diets

were mashed, sieved through a 200–800 micron sieve, to

get 0.2–0.8 mm pellets, labelled and stored at
20 °C until

used

The diets were initially fed to the larvae at a daily rate

of 100 g kg
1of their body weight (BW) The feeding level

was reduced to 70 g kg
1 BW beginning of the second

month The diets were offered three times a day (08:00,

12:00 and 16:00 h), 7 days a week, for 60 days Fish were

weighed collectively at 15-day intervals, their average

weights were recorded, and the daily amount of feed for

each aquarium was readjusted accordingly

A 350 g kg
1 crude protein diet (Table 2) was lated and used for feeding Nile tilapia fingerlings inexperiment 2 Diet preparation, processing and colour-ation were similar to that used in experiment 1 Thesame diet colours used in experiment 1 were alsoadopted in Experiment 2 The diets were manually fed tothe fish to satiation, three times a day (08:00, 12:00 and16:00 h), 7 days a week, for 60 days Fish were weighedcollectively at 15-day intervals and their average weightswere recorded

formu-At the end of each experiment, fish in each aquarium werenetted, counted, weighed and frozen at
20 °C for finalbody composition analyses Initial body analyses wereperformed on a pooled sample of fish, which was weighedand frozen before each experiment A sample of each testdiet was also stored
20 °C for chemical analysis Proxi-mate analyses of whole-body moisture, protein, lipid andash were performed according to standard AOAC (1995)methods

Growth rates and feed efficiency were calculated as follows:

Per cent weight gainðPWGÞ ¼ 100ðWf
WiÞ=WiSpecific growth rateðSGRÞ ¼ 100ðLnWf
LnWiÞ=t

Where Wi and Wf are initial and final weights (g), and t

is time of experiment (days)

Feed conversion ratio (FCR) = dry feed intake (g)/fishlive weight gain (g)

Protein productive value (PPV)= 100(protein gain (g)/

protein fed (g))

All data were subjected to one-way analysis of variance(ANOVA) at a 95% confidence limit, usingSPSSsoftware, ver-sion 12 Duncan’s Multiple Range test was used to com-pare means when F-values from theANOVAwere significant(P< 0.05)

Table 2 Formulation and proximate analysis of the tested diets for

Nile tilapia O niloticus larvae (Experiment 1) and fingerlings

1 700 g kg
1crude protein (Denmark).

2 680 g kg
1crude protein (Morocco).

3 Contains (kg
1): Vit A, 15000 I.U.; Vit D3, 1500 I.U.; Vit E,

2.0 mg; Vit K3, 2.0 mg; Riboflavin (B2), 2.5 mg;

Calcium-D-Panto-thenate, 5.5 mg; Nicotinamide (B3), 10.0 mg; Pyridoxine HCl (B6),

3.0 mg; Thiamine HCl (B1), 2.0 mg; Vit B12, 5.0 mg; Folic acid,

2.0 mg; Niacin, 1.0 mg; Mn, 60.0 g; Fe, 30.0 g; Cu, 4.0 g; Zn,

50.0 g; I, 300 mg; Co, 100 mg; Se, 100 mg; CaCO3, 855.5 g.

4 Carboxymethyl cellulose, used as binder.

5 Nitrogen-free extract, determined by difference.

6 Gross energy, calculated based on 23.64, 39.54 and 17.57

(KJ g
1) for protein, lipid, carbohydrate, respectively.

.

The results of experiment 1 indicated that feed colours

sig-nificantly affected (P< 0.05) the growth rates and feed

uti-lization of Nile tilapia O niloticus larvae (Table 3, Fig 1)

Dark-coloured diets (dark blue, dark brown, red and dark

green, respectively) provided the best performance

(P < 0.05), while light-coloured diets (yellow and light

brown) resulted in a sharp reduction in larval performance

Fish survival was also significantly affected by diet colours

(P < 0.05) Dark-blue and dark-green diets produced

signif-icantly higher survival (95% and 87%, respectively) than

did the lighter-coloured diets (yellow and light brown)

(63% and 55%, respectively)

Body composition of Nile tilapia O niloticus larvae was

significantly affected (P< 0.05) by diet colour (Table 4)

Dark-coloured diets resulted in the lowest body water

con-tent and the highest protein concon-tent, while lighter-coloured

diets (light brown and yellow) provided the highest body

water content and lowest protein content Body lipid and

ash contents were also significantly affected (P < 0.05) by

diet colours, but they showed irregular patterns

The results of experiment 2 indicated that the growth rates

and survival of O niloticus fingerlings were not

signifi-cantly affected (P> 0.05) by feed colours (Table 5) Feed

colours also did not significantly affect feed utilization

effi-ciency Body composition of Nile tilapia fingerlings fed the

test diets was also not significantly affected by diet colours

(P > 0.05) (data not presented)

The onset of exogenous feeding is a bottle-neck stage in

larval development of fish Some fish species are visual

feeders during their larval stages, where food colour

enhances food capture efficiency (El-Sayed & El-Ghobashy2011) Other species may respond equally to both visualand chemical stimuli at the start of exogenous feeding(Valente et al 2001) Feed colour preferences may alsodiffer depending upon culture conditions, such as tankcolour and light intensity (Ginetz & Larkin 1973; El-Sayed

& El-Ghobashy 2011) and pellet density, size, colour andtexture (El-Sayed 2006)

Little information is available on the effects of feed our on the performance of tilapia, particularly during theirearly life stages In the present study, dark-coloured dietsprovided better growth performance and feed efficiency offish larvae than light-coloured diets, presumably due to thesharp contrast between these diets and the background col-ours (light colour) Fish larvae fed on the light-coloureddiets also have significantly higher mortality rates Thismight have been due to the low contrast between thesediets and tank background, which may have reduced fishability to visualize, detect and ingest the feed particles, and

col-in turn, they had to spend more energy searchcol-ing for food

Table 3 Growth, feed utilization and survival (mean  SEM) of Nile tilapia larvae fed the test diets in experiment 1

Light brown (control) 0.22  0.1 a 2264  16 a 5.24  0.36 a 1.69  0.14 a 21.67  2.27 a 55.00  2.89 a Dark Blue 0.64  0.0 b 6715  24 b 7.03  0.21 b 2.06  0.14 b 30.11  1.10 b 95.00  2.89 b Dark Green 0.54  0.0 c 5654  18 c 6.75  0.13 b 1.78  0.09 a 28.48  1.50 b 86.67  3.33 c Dark Brown 0.56  0.1 c 5854  37 c 6.81  0.23 b 1.97  0.12 b 26.04  1.16 c 70.00  5.77 d Red 0.55  0.1 c 5787  23 c 6.79  0.15 b 2.09  0.13 b 22.53  1.34 a 71.67  4.41 d Yellow 0.32  0.0 d 3314  39 d 5.87  0.38 a 1.81  0.15 a 24.39  2.13 c 63.33  6.00 e Values in the same column with different superscripts are significantly different at P < 0.05.

0 0.1 0.2 0.3 0.4 0.5 0.6

Figure 1 Growth rates of Nile tilapia larvae fed the test diets in experiment 1.

Despite that high mortality may affect (improve) fish

growth, by providing the surviving fish with more space for

growth, this is not the case in our study, because the fish

that had the highest mortality exhibited the lowest growth

rates In general, the present results suggest that Nile

tila-pia O niloticus larvae are visual feeders at the onset of

exogenous feeding

Similar results have been reported on thinlip mullet (Liza

ramada) fry, where fish fed the dark-coloured diets had

bet-ter growth, feed efficiency and survival than those fed

lighter diets (El-Sayed & El-Ghobashy 2011) Ginetz &

Larkin (1973) found also that rainbow trout Salmo

gaird-neri was more attracted to darker-coloured prey than to

lighter-coloured prey, in combination with a pale

greenish-blue background Similarly, Browman & Marcotte (1987)

found that more Atlantic salmon alevin consumed more

red-coloured prey (copepods) against blue background than

against red or green background, while blue-coloured

cope-pods were more consumed against green and blue

back-grounds These feed preferences have been attributed to the

contrast between the feed (dark) and the background

(light) (Endler 1991; Utne-Palm 1999) This contrast leads

to higher visibility of the food and better ability of young

fish to perceive food items and, in turn, leads to higher

feeding efficiency, growth rates and survival

The visual-feeding behaviour emphasizes the role of

vision sense in detection, distinguishing and consumption

of food items during fish early life stages (Downing &

Litvak 1999; Martinez-Cardenas & Purser 2007) Thisbehaviour enables better prey visualization and subse-quently improved feed intake (Martin-Robichaud & Peter-son 1998) However, the ability of visual-feeding fish todetect and ingest feed can also be affected, among otherfactors, by physical characteristics of the feed, such as pel-let density (sinking versus floating), size, colour, shape andtexture (El-Sayed 2006)

The visual-feeding behaviour of Nile tilapia larvae in thepresent study could be related to the development of eyeretina, which plays a significant role in colour vision andfood detection The retina of most fish larvae at hatchingconsists mainly of cones, whereas rods develop at laterstages (O’Connell 1981; Blaxter 1986) However, moreresearch is needed to identify the maximum cone absor-bance that drives visual acuity in Nile tilapia larvae, andthe visual spectrum at which they are able to best detectfood colours The appearance of rod cells also increaseslarval photosensitivity, enables the fish to discriminate col-ours and suggests that changes in lighting regimes could benecessary throughout the larval phase (Nicol 1963) Forexample, Lythgoe & Shand (1989) reported that photopicvision is usually mediated by two to four cone types con-taining different visual pigments which allow the possibility

of colour vision

The preference variations between different colours inthe present study may also be related to the representa-tion of these colours in fish brain, which can discriminate

Table 5 Growth rate and feed utilization (mean  SEM) of Nile tilapia fingerlings fed the test diets in experiment 2

Light brown (Control) 5.32  0.05 25.97  1.33 a

Yellow 5.30  0.06 25.60  1.54 a 383.0  14.6 a 2.62  0.05 a 0.88  0.06 a 37.92  2.22 a 100.0  0.00 a

Values in the same column with different superscripts are significantly different at P < 0.05.

Table 4 Body composition (g kg ) (mean  SEM) on wet weight basis of Nile tilapia larvae fed the test diets in experiment 1

colours (Nicol 1963; Lythgoe & Shand 1989) This may

have been exposed clearly in case of dark-coloured diets

(particularly dark blue and dark green diets), which may be

translated as a natural food colour in the wild

environ-ment, or as a comfortable colour for fish performance, as

suggested by Volpato & Barreto (2001) and Volpato et al

(2004) Consequently, dark-coloured diets could be the

low-est stressing for fish, lowlow-est energy demanding, and, in

turn, the highest growth promoters Nonetheless, more

research is needed to support this assumption

On the contrary of the above results, Arumugam (1997)

reported tilapia hybrid larvae showed no preference in

terms of diet colour when they were fed three coloured

encapsulated diets (red, orange and green) However, that

study was carried out for a very short period (<6 days) If

the fish had been exposed to these diets for longer periods,

and more colours were tested, those results would have

been more reliable

In the present study, all the test diets offered to Nile

tila-pia fingerlings had the same composition, but had different

colours Therefore, the non-significant effects of diets

col-ours on fish performance suggest that Nile tilapia

finger-lings are not visual feeders, and feed colour may not be an

important factor in feed detection and consumption, as

compared to larval stages Other non-visual stimulators,

including feed taste, odour, size, form and palatability may

be more important for food detection during fingerling and

adult stages than vision sense (Villamizar et al 2009)

Several studies have supported this argument and

indi-cated that other physiological processes, apart from the

vision–feeding relationship, play more roles in feeding

behaviour of juvenile and adult fish For example, Kallayil

et al (2003) found that foraging behaviour of cod was

chemically stimulated by bait odour, even under non-visual

conditions Batty & Hoyt (1995) found also that juvenile

sole and plaice were able to feed on dead prey when only

chemical stimuli were available However, sole relied

princi-pally on chemoreception and mechanoreception, where

pla-ice showed a greater dependence on vision for feeding It

has also been reported that juvenile alewife (Alosa

pseud-oharengus) feed in nature on zooplankton (such as Mysis

relicta, Daphnia magna and Artemia salina) at night

(Jans-sen et al 1995) This behaviour has been attributed to the

lateral line, which was used by the fish to sense the prey

and feed efficiently in the dark In a recent study,

Hirt-Chabbert et al (2012) found that diets containing two

stimulants (processed marine proteins and yeast protein)

provided better growth, homogenous size distribution and

feed intake in elver stages of European eels (Anguilla

angu-illa) than the control group On the other hand, these ulants did not significantly affect the growth performanceand size distribution of glass eels

stim-In some fish species, food search response is generallyprovided simultaneously by both the olfactory and theexternal gustatory systems in a complimentary manner Insupport, Kasumyan & Marusov (2007) found that the abil-ity of the fish to recover sensitivity to food smells after aprolonged olfactory deprivation is provided by compensa-tory processes taking place in the external gustatory sys-tem They also suggested that the complementarity ofchemosensory systems, their functional interaction andcapacity for a compensatory development can be consid-ered as a sensory mechanism providing the reliability ofrealization of feeding behaviour by fish

In conclusion, the present study revealed that diet colour

is an important factor that should be considered in tilapiaculture, especially during their early larval stages Dark-col-oured diets (particularly dark-blue diet) are more preferable

by O niloticus larvae for optimum growth performanceand survival The study also suggested that fingerling Niletilapia may not be visual feeders, and other non-visualstimulators are more important for food detection thanvision sense However, additional, long-term studies should

be conducted using other size classes of cultured Nile pia to support the present findings

tila-Arumugam, P.T (1997) Suitability of a continuous-flow chamber for investigating fish larvae/fry growth responses Aquaculture,

151, 365–370.

Association of Official Analytical Chemists (AOAC) (1995) Official Methods of Analysis, 16th edn, AOAC, Arlington, VA, USA Batty, R.S & Hoyt, R.D (1995) The role of sense organs in the feeding behaviour of juvenile sole and plaice J Fish Biol., 47,

931 –939.

Biswas, A.K & Takeuchi, T (2002) Effect of different photoperiod cycles on metabolic rate and energy loss of both fed and unfed adult tilapia Oreochromis niloticus: part II Fish Sci., 68, 543– 553.

Blaxter, J.H.S (1986) Development of sense organs and behaviour

of teleost larvae with special reference to feeding and predator avoidance Trans Am Fish Soc., 115, 98–114.

Browman, H.I & Marcotte, B.M (1987) Effect of prey color and background color on feeding by Atlantic salmon alevins Prog Fish Cult., 49, 140–143.

Downing, G & Litvak, M.K (1999) The effect of photoperiod, tank colour and light intensity on growth of larval haddock Aquacult Int., 7, 369–382.

El-Sayed, A.-F.M (2004) Feed colour affects growth, feed tion of Nile tilapia Global Aquacult Advoc., 7, 76.

utiliza-El-Sayed, A.-F.M (2006) Tilapia Culture CABI publishing, CABI International, Willingford, Oxfordshire, UK, 274 pp.

El-Sayed, A.-F.M (2007) Analysis of feeds and fertilizers for

sus-tainable aquaculture development in Egypt In: Study and

Anal-ysis of Feeds and Fertilizers for Sustainable Aquaculture

Development (Hasan, M.R ed.), pp 401 –422 FAO Fish Tech.

Pap No 497, FAO, Rome, Italy.

El-Sayed, A.-F.M & El-Ghobashy, A.E (2011) Effects of tank

colour and feed colour on growth and feed utilization of

thinlip mullet (Liza ramada) larvae Aquacult Res., 42, 1163–

1169.

El-Sayed, A.-F.M & Kawanna, M (2004) Effects of photoperiod

on the performance of farmed Nile tilapia Oreochromis niloticus:

I Growth, feed utilization efficiency and survival of fry and

fin-gerlings Aquaculture, 231, 393–402.

Endler, J.A (1991) Interactions between predators and prey In:

Behavioural Ecology: An Evolutionary Approach, 3rd edn

(Krebs, J.R & Davies, N.B., eds.), pp 169 –196 Blackwell,

Oxford, UK.

Fanta, E (1995) Influence of background colour on the behaviour

of the fish Oreochromis niloticus (Cichlidae) Arq Biol Tecnol.,

38, 1237–1251.

Fiksen, Ø., Utne, A.C.W., Aksnes, D.L., Eiane, K., Helvik, J.V &

Sundby, S (1998) Modelling the influence of light, turbulence

and ontogeny on ingestion rates in larval cod and herring Fish.

Oceanogr., 7, 355–363.

Food and Agriculture Organization of the United Nations (FAO)

(2012) Global Aquaculture Production 1950 –2010 FAO, Rome,

Italy

(http://www.fao.org/fishery/statistics/global-aquacul-ture-production/en).

Fortes-Silva, R., Mart ınez, F.J & Sanchez-Vazquez, F.J (2011)

Macronutrient selection in Nile tilapia fed gelatin capsules and

challenged with protein dilution/restriction Physiol Behav., 102,

356 –360.

Ginetz, R.M & Larkin, P.A (1973) Choice of colours of food

items by rainbow trout (Oncorhynchus mykiss) J Fish Res Bd.

Can., 30, 229–234.

Gro¨neveld, D., Balm, P.H.M., Martens, G.J.M., & Wendelaar

Bonga, S.E.W (1995) Differential melanin-concentrating

hor-mone gene-expression in two hypothalamic nuclei of the teleost

Tilapia in response to environmental changes J Neuroendocrinol.,

7, 527–533.

Hirt-Chabbert, J.A., Skalli, A., Young, O.A & Gisbert, E (2012)

Effects of feeding stimulants on the feed consumption, growth

and survival at glass eel and elver stages in the European eel

(Anguilla anguilla) Aquac Nutr., 18, 152–166.

Janssen, J., Jones, W.R., Whang, A., Philip, E & Oshel, P.E.

(1995) Use of the lateral line in particulate feeding in the dark

by juvenile alewife (Alosa pseudoharengus) Can J Fish Aquat.

Sci., 52, 358–363.

Jegede, T & Olusola, O.D (2010) Effects of feed colour on growth

and nutrient utilization of Tilapia zillii and Oreochromis niloticus

Fingerlings Agric Biol J N Am., 1, 1182–1186.

Kallayil, J.K., JØrgensen, T., Engas, A & Ferno, A (2003)

Baiting gill nets- how is fish behaviour affected? Fish Res., 61,

125 –133.

Kasumyan, A.O & Marusov, E.A (2007) Chemoreception in chronically anosmiated fish: a phenomenon of compensatory development of the gustatory system J Ichthyol., 47, 647–655.

Luchiari, A.C., Duarte, C.R.A., Freire, F.A.M & Nissinen, K.

(2007) Hierarchical status and colour preference in Nile tilapia (Oreochromis niloticus) J Ethol., 25, 169–175.

Lythgoe, J.N & Shand, J (1989) The structural basis for iridescent colour changes in dermal and corneal iridophores in fish J Exp.

Biol., 141, 313–325.

Martinez-Cardenas, L & Purser, G.J (2007) Effect of tank colour

on Artemia ingestion, growth and survival in cultured early nile pot-bellied seahorses (Hippocampus abdominalis) Aquacul- ture, 264, 92–100.

juve-Martin-Robichaud, D.J & Peterson, R.H (1998) Effects of light intensity, tank colour and photoperiod on swimbladder inflation success in larval striped bass, Morone saxatilis (Walbaum).

Rad, F., Bozaoglu, S., Ergene Gozukara, S., Karahan, A & Kurt,

G (2006) Effects of different long-day photoperiods on somatic growth and gonadal development in Nile tilapia (Oreochromis niloticus L.) Aquaculture, 255, 292–300.

Santiago, C.B., Aldaba, M.B & Reyes, O.S (1987) Influence of feeding rate and diet form on growth and survival of Nile tilapia (Oreochromis niloticus) fry Aquaculture, 64, 277–282.

Turker, H., Eversole, A.G & Brune, D.E (2003) Effect of Nile tilapia, Oreochromis niloticus (L.), size on phytoplankton filtra- tion rate Aquacult Res., 34, 1087–1091.

Utne-Palm, A.C (1999) The effect of prey mobility, prey contrast, turbidity and spectral composition on the reaction distance of Gobiusculus flavescens to its planktonic prey J Fish Biol., 54,

1244 –1258.

Valente, L.M.P., Saglio, P., Cunha, L.M & Fauconneau, B (2001) Feeding behaviour of fast- and slow-growing strains of rainbow trout, Oncorhynchus mykiss (Walbaum), during first feeding.

Aquacult Res., 32, 471–480.

Van der Salm, A.L., Metz, J.R., Bonga, S.E & Flik, G (2005) Alpha-MSH, the melanocortin-1 receptor and background adap- tation in the Mozambique tilapia, Oreochromis mossambicus.

Gen Comp Endocrinol., 144, 140–149.

Villamizar, N., Garc ıa-Alcazar, A & Sanchez-Vazquez, F.J (2009) Effect of light spectrum and photoperiod on the growth, devel- opment and survival of European sea bass (Dicentrarchus labrax) larvae Aquaculture, 292, 80–86.

Volpato, G.L & Barreto, R.E (2001) Environmental blue light prevents stress in the fish Nile tilapia Braz J Med Biol Res.,

1 1 1 2 2 2 3

1 Institute of Oceanography and Fisheries, Split, Croatia;2 Department of Marine Studies, University of Split, Split, Croatia;

3

Ichthyopathology Group – Biological Materials, Division of Material Chemistry, Ruder Bo
skovic Institute, Zagreb,Croatia

To study the effect of propolis in crude form (CPP),

prepared without any chemical refinement (CPP) on

Dicen-trarchus labraxunder low-temperature stress, sea bass

juve-niles were randomly divided into three groups: a control

group fed with basal diet and two treatment groups fed

with basal diet supplemented with 1.25 and 2.5 g kg 1 of

propolis At the end of a 10-week feeding trial, sea bass

were exposed to low-temperature stress at 12 °C for 24 h

The growth performance, RNA/DNA ratio and changes in

serum biochemical parameters were investigated Dietary

intake of propolis stimulated the specific growth rate

(SGR), feed conversion efficiency (FCE), RNA/DNA ratio

and alkaline phosphatase (ALP) enzyme activity, while

decreasing plasma triglycerides and aspartate

aminotrans-ferase (AST) activity Supplement of 2.5 g kg 1 CPP in

diet significantly increased the mean SGR and FCE up to

9% and 13.4%, respectively, in comparison with the

con-trol group Low-temperature stress elevated serum

triglyce-rides, glucose and cortisol levels in all groups; however,

glucose and cortisol reached significantly lower end values

in group receiving highest amount of propolis in diet This

study suggests that ingestion of basal diet supplemented

with 2.5 g kg 1of propolis could prevent adverse effects of

low-temperature stress and promote the growth of sea

bass

KEY WORDS: biochemical parameters, Dicentrarchus labrax,

growth, propolis diet, RNA/DNA ratio

Received 18 September 2012; accepted 26 November 2012

Correspondence: Tanja
Segvi c-Bubic, Institute of Oceanography and Fisheries, PO Box 500,
Setali
ste Ivana Me
strovica 63, 21000 Split, Croatia E-mail: tanja.segvic@izor.hr

It is well known that fish in culture are exposed to varioushusbandry-related acute (e.g handling, temperature varia-tion) and chronic (e.g poor water quality, confinement)stressors (Barton 2002) Both adversely affect fish physiol-ogy and behaviour, often reflected in growth rate reduction(Vijayan & Leatherland 1988), as well as immunosuppres-sion (Ndong et al 2007), leading to increased susceptibility

to various pathogens As a control measure, fish feed iscommonly supplemented with large number of antibioticsand hormones, resulting in drug resistance, residues andother issues detrimental to the sustainable development ofaquaculture, seafood safety and human health Therefore,using naturally derived additives instead, like medicinalplants (D€ugenci et al 2003), probiotics (Bandyopadhyay &Mohapatra 2009) and beehive products (Cuesta et al 2005)would be highly favourable These non-specific immuno-stimulants have recently received increasing attention asthey can be easily and cost-effectively incorporated into thediet and elicit low environmental impact (Cuesta et al.2005)

Propolis is a resinous hive product collected by bees from plants, showing a very complex chemical compo-sition It has been used in folk medicine since ancienttimes, due to its many biological properties, such as anti-bacterial (Sforcin et al 2000), antitumor (Banskota et al

.

Aquaculture Nutrition

2002; Bazo et al 2002) and immunomodulatory (Murad

et al.2002), among others Analyses of its chemical

compo-sition have identified at least 300 compounds (De Castro

2001) Biological activities of propolis are mainly attributed

to the phenolic components such as flavonoids Flavonoids

have been reported to exhibit a wide range of biological

activities, including antibacterial, antiviral,

anti-inflamma-tory, antiallergic and vasodilatory actions In addition,

flavonoids inhibit lipid peroxidation, platelet aggregation,

capillary permeability and fragility, and the activity of

enzyme systems including cyclooxygenase and lipoxygenase

(Havsteen 2002)

Recently, the effects of propolis as a growth promoter

or/and as an immunostimulant have been extensively

dem-onstrated in poultry and fish (Cuesta et al 2005; Denli

et al 2005; Chu 2006; Abd-El-Rhman 2009; Meurer et al

2009; Talas & Gulhan 2009; Cß etin et al 2010; Deng et al

2011) However, the benefits of propolis use on European

sea bass (Dicentrarchus labrax) physiology have not yet

been reported The European sea bass is a marine,

eury-haline teleost with important socio-economic value for the

Mediterranean aquaculture industry, prompting vivid

sci-entific interest over the last 35 years Nevertheless, main

impingement to sea bass culture remains the

aquaculture-related stress itself Varsamos et al (2006) pointed out

that although within the normal temperature range for

such a poikilothermic species, sudden shifts in temperature

can be harmful The objective of this study was to

evalu-ate the effect of dietary intake of crude propolis powder

(CPP) on growth performance and low-temperature stress

resistance in sea bass juveniles by measuring serum

bio-chemical parameters: cortisol (COR), glucose (GLU),

tri-glycerides (TRIG), cholesterol (CHOL), aspartate

aminotransferase (AST), alkaline phosphatase (ALP) and

total protein (TP)

All fish handling procedures followed established standards

for the human care and use of animals during sampling

period, which were previously approved by the Ethical

Committee for Animal Welfare of the Institute of

Oceanog-raphy and Fisheries, Croatia

Propolis was obtained from a farm in Dalmacija region,

middle Croatia To pulverize the gluey crude propolis and

prepare it for final usage, keeping all propolis substances,

an original method of treatment without any chemical cedures by firm HEDERA d.o.o (Stobrec, Croatia) wasused (Sobocˇanec et al 2011) The normal pellet diet AllerFutura (Aller Aqua, Denmark, 640 g kg 1 crude protein,

pro-120 g kg 1crude fat, 50 g kg 1carbohydrates, 110 g kg 1

ash as declared by the producer) was crushed, mixed withtap water containing the adequate amount of propolispowder and made again into pellets, thus obtaining dietssupplemented with 0 (control), 1.25 or 2.5 g kg 1of propo-lis Remade pellets were allowed to dry and stored at 4°C

Diets were fed daily at a rate of 3% per fish

European sea bass juveniles (70-day posthatch (DPH))were obtained from commercial hatchery Cenmar d.d

(Nin, Croatia) and transferred to the Laboratory of culture, Institute of Oceanography and Fisheries, Split Thejuveniles were raised at 23 0.5 °C and fed a commercialsea bass dry formulated diet Aller Futura (Aller Aqua,Denmark), for 3 weeks prior to starting the feeding trial

Aqua-After acclimation, juveniles were measured, weighted andrandomly distributed into 9 tanks (groups of 60 fish) withthree replicates per dietary treatment The initial meanweight (SE) was 1.60  0.03 g, and it did not differ sig-nificantly among experimental units Experiment lasted

10 weeks, and food ration was adjusted after day 40 lowing weighing and measuring Dietary treatment included

fol-a control group fed with bfol-asfol-al diet fol-and two groups fedwith basal diet supplemented with 1.25 and 2.5 g kg 1 ofcrude propolis powder, that is, 1.25 g kg 1 CPP and2.5 g kg 1 CPP group, respectively On 69th day of theexperiment, low-temperature stress regime was preformed:

temperature was rapidly lowered at 12°C with a decrease

in rate of 1°C/15 min and maintained at 12 °C for 24 h

The feeding trial was conducted in 70-L plastic tanks in

an indoor air-conditioned closed recirculation culture tem supplied with continuously filtrated sea water(15 L min 1) at a temperature of 23 0.5 °C Tempera-ture, salinity and dissolved oxygen were monitored daily

sys-Fish were hand-fed every hour at 3% BW day 1, havingartificial day cycle of 8 h (from 8 am until 16 pm) Duringthe experiment, care was taken to feed the fish to apparentsatiation, that is, feeding was halted when uneaten foodpellets were observed in the tank bottom The fish was notfed the day before sampling, on sampling days and notduring the cold-stress day No mortality was observed dur-ing the experiment

.

At the end of the experiment, all fish were anaesthetized

with benzocaine (4% in acetone), weighted and measured

individually Simultaneously, white muscle and blood were

extracted from eight different and randomly chosen fish in

each experimental tank to analyse RNA/DNA ratio and

blood physiology, respectively Muscle samples were

imme-diately frozen in liquid nitrogen and stored at 80 °C until

analyses

After the dietary intake of propolis, the specific growth rate

(SGR,% body weight/day) for each tank was determined

using the equation SGR= 100(lognWf–lognW0)/t, where

W0and Wfwere the initial and final average weights (g) of

each experimental tank, respectively, after t days (Ricker

1979) Feed conversion efficiency (FCE) was calculated

using the equation FCE= (B2–B1)/C, where B2and B1are

biomass in tank (g) at days t2 and t1, respectively, and C is

feed consumption in the tank during the period

Growth is sustained by intensive protein synthesis, and

cel-lular RNA level varies accordingly As DNA content

remains constant, RNA/DNA ratio points to physiological

and nutritional state of an organism It has been proven as

a reliable indicator of recent growth and nutritional status

in juvenile fish (Buckley et al 1999; Gwak et al 2003) as

well as in D labrax larvae (Berdalet et al 2005b)

RNA/DNA ratio of the white muscle was determined

using a fluorescent dye-binding method as described in

Berdalet et al (2005a) Nucleic acids were extracted

follow-ing the procedure compatible with this protocol (Berdalet

et al 2005b) In the RNA/DNA assay, the nucleic acid

content is measured as fluorescence intensity of bound

SYBR Green II fluorochrome (Molecular probes;

excita-tion 497 nm, emission 520 nm), and specific contribuexcita-tion

of RNA and DNA was determined by nuclease treatment

of the sample with DNase (Sigma D4263) and RNase

(Sigma R6513), respectively A third assay is

simulta-neously run with both nucleases to account for residual

flu-orescence

The protocol was adapted for a 96-well microplate

for-mat with the assay final volume set at 200 lL, and to

mini-mize pipetting errors, concentrations of working solutions

were adjusted to allow component addition in minimal

increments of 20lL The standard curves were run

con-comitantly with the sample readings: we used genomicDNA from calf thymus (Sigma D4764) as DNA standardand 16S/23S ribosomal RNA from Escherichia coli (Roche10206938001) as RNA standard To allow comparability ofRNA/DNA ratio results between laboratories, we reportstandard curve slope ratio mDNA/mRNA of 1.25 asdescribed in Berdalet et al (2005b) and Caldarone et al.(2006)

To monitor possible synergistic effect of different lis-containing diets and consequently cold-stress regime onblood physiology, blood samples (0.1–0.5 mL) wereextracted with 1-mL syringes from the caudal vessels ofeight fish from each experimental tank the day beforeand the day after the cold-stress treatment Samples werecollected in tubes coated with anticoagulant lithium hepa-rin, centrifuged at 12 000 g for 90 s and the resultantplasma was frozen at 20 °C for storage until the analy-sis 1 month later The determination of biochemicalparameters total proteins (TP), triglyceride (TRIG),plasma cholesterol (CHOL), glucose (GLU), alkalinephosphatase (ALP) and aspartate aminotransferase (AST)were carried out using a SABA-18 autobiochemistry ana-lyser (AMS Analyzer Medical System, Rome, Italy) andspectrophotometrically using the BioMerieux (Lyon,France) diagnostic kit

propo-To extract the serum for the determination of cortisol(COR), the remaining fraction of blood samples not trea-ted with heparin was allowed to clot at 4°C, centrifuged

at 1500 g for 10 min and stored at 80°C until analysis.Serum cortisol concentration was determined with com-mercially available enzyme-linked (ELISA) immunoassaykit (Alpha Diagnostic International, San Antonio, TX,USA) Validation of its specificity in recognizing the fishcortisol molecule has already been tested by Caruso et al.(2011)

To assess normality of distributions, a Kolmogorovnov test (Zar 1984) was used and homogeneity of variancestested using Levene’s F-test Analysis of mean weight andcondition on the three sampling dates was conducted using

–Smir-a two-w–Smir-ay nested ANOVA, where the replicates (random)were nested within propolis-containing diets (fixed) Forparameters where only group data existed (SGR, FCE andsurvival), a one-way ANOVAwas applied Two-way analysis

of covariance (ANCOVA) was performed to determine the

effect of cold stressor, propolis-containing diet and their

interaction on serum physiological parameters where the

fish weight was used as a covariate One-way ANCOVA was

applied to detect any difference caused by dietary

treat-ments at fixed sampling time (before/after the cold stress)

and/or by time sampling at fixed dietary treatment, as well

as to determine the effect of propolis-containing diet on

RNA/DNA ratio SignificantANOVAandANCOVA’s were

fol-lowed by a Student–Newman–Keuls multiple comparison

test (Zar 1984) to identify differences among treatments

A significance level (a) of 0.05 was used if not stated

otherwise

Mean body weight (W) and total length (L) for every

group and measurement are summarized in Table 1, as well

as the analyses of differences in W and L among groups

At the end of both measuring periods, 2.5 g kg 1 CPP

group was longer and heavier in comparison with the rest

of the groups (P < 0.05), while no differences were

observed between the 1.25 g kg 1CPP group and the

con-trol group (P> 0.05) The final mean weight was 15–18%

larger in the 2.5 g kg 1CPP group compared with the

con-trol group Survival of fish ranged from 98.3 to 99.3% and

was not significantly different among the dietary treatments

(Table 2) Dietary supplementation with 2.5 g kg 1 CPP

significantly improved the specific growth of fish and feed

efficiency ratio, but no significant differences were observed

between the 1.25 g kg 1 CPP and the control group

(Table 2) In addition, RNA/DNA ratio in muscle was

sig-nificantly higher in the 2.5 g kg 1 CPP group in

compari-son with the 1.25 g kg 1 CPP and the control group

(Table 2)

The effects of different dietary treatments and cold-shockregime performed 24 h before the final sampling, observedthrough serum biochemical parameters, are summarized inFig 1 Overall, dietary treatment had significant impact onserum triglycerides, ALP and AST, while cold stress signifi-cantly affected serum triglycerides, glucose and cortisol lev-els, with the latter two in diet-dependent manner (two-way

ANCOVA, Table 3) Before the cold stress, both groups ofsea bass juveniles fed with propolis-enriched diet showedmarkedly lower levels of serum triglycerides than the con-trol group (P< 0.05), while levels of ALP and AST signifi-cantly changed only in 2.5 g kg 1 CPP-supplementedgroup: the increase in plasma ALP and the decrease inAST were observed After the cold stress, triglyceridesincreased in all diet groups (P< 0.05) reaching similar endvalues (P> 0.05), whereas AST and ALP levels remainedunaffected (P> 0.05), conserving general trends as previ-ously observed Exposure to temperature drop furtherincreased glucose levels in all groups where significantlylower level was observed for the 2.5 g kg 1 CPP as com-pared to the control group, although the same was not truefor the 1.25 g kg 1 CPP group In addition, all groupsexhibited a significant increase in serum cortisol levels,however, less pronounced in groups that received dietaryintake of propolis as compared to the control group(P< 0.05) No significant differences were found in theplasma total protein and cholesterol levels among groupsbefore and after the cold-stress treatment

Apart from the studies of propolis impact on the immunesystem and fish disease resistance (Cuesta et al 2005; Chu2006; Abd-El-Rhman 2009; Zhang et al 2009), in recentyears, great number of studies are carried out on propolis

Table 1 Body weight (W) and total length (L) of European sea bass Dicentrarchus labrax juveniles fed with experimental diets containing

different amounts of propolis, measured at the beginning of the experiment, at the day 40 and at the end of study period (day 70)

Propolis concentration (g kg 1 diet)

within each measurement date (Student–Newman–Keuls test, P < 0.05).

.

Table 2 Mean specific growth rate (SGR), feed conversion efficiency (FCE) and RNA/DNA ratio of European sea bass Dicentrarchus rax juveniles fed with experimental diet containing different amounts of propolis for 70 days

lab-Propolis concentration (g kg 1 diet) Survival (%) SGR (% day 1 ) FCE RNA/DNA ratio 1

1 Comparison based on ANCOVA test using final body weight as covariate.

Figure 1 Effect of dietary supplementation with different levels of crude propolis powder (CPP) on serum biochemical parameters of pean sea bass Dicentrarchus labrax fingerlings subjected to an acute cold stressor Values are given as means of three replicates SEs Dif- ferent capital letters above the bars indicate significant differences (P < 0.05) at different sampling points of the same group, and different small letters above the bars indicate significant differences in different groups of the same sampling points in Student –Newman–Keuls test.

Euro-metabolism and consequently on fish growth performances

(Meurer et al 2009; Deng et al 2011; Bae et al 2012) The

use of immunostimulants as dietary supplements can

improve the innate defence of animals providing resistance

to pathogens during periods of high stress, such as grading,

reproduction, sea transfer and vaccination (Bricknell &

Dalmo 2005) Nevertheless, studies focused on

Mediterra-nean farmed finfish, and the use of propolis as

immuno-stimulant and growth promoter is scarce

This study shows that dietary 2.5 g kg 1CPP

supplemen-tation can improve European sea bass growth, feed

conver-sion efficiency and can reduce physiological acute stress

response, pinpointing the benefits of propolis use as a feed

additive Moreover, dietary supplementation of 2.5 g kg 1

CPP increased significantly the mean SGR and FCE up to

9% and 13.4%, respectively, in comparison with the control

group Similarly to our results, studies with rainbow trout

(Oncorhynchus mykiss) and juvenile eel (Anguilla japonica)

showed that incorporation of 2–4 g kg 1

(Deng et al 2011)and 5 g kg 1(Bae et al 2012) of ethanolic extract of propo-

lis increased the growth performance and decreased the feed

conversion ratio In addition, experiments with Nile tilapia

(Oreochromis niloticus) showed that the diet supplemented

with 1.83–2.74 g kg 1

of brown propolis extract (Meurer

et al.2009) and 10 g kg 1of ethanolic extract of propolis or

crude propolis (Abd-El-Rhman 2009) markedly increased

the growth performance These studies suggested that

benefi-cial effect of propolis on fish growth arises due to the

antimi-crobial activity of the propolis extract components

(Kujumgiev et al 1999), resulting in better intestinal healthand improved digestion and absorption In addition to itsantimicrobial properties, propolis possesses antioxidant bio-logical activities (Orhan et al 1999) In contrast, few studies(Velotto et al 2010; Kashkooli et al 2011) pointed out nopositive propolis effect on fish weight gain and specificgrowth rate, although mortality of fish eggs was reduced andmuscular growth improved These different results may bedue to the dose used and/or different propolis origin Propo-lis composition varies with different factors, such as thesource of the exudates, climate and environmental condition(Chen & Wong 1996; Nieva-Moreno et al 1999), namelyseveral researchers (Cowan 1999; Chen et al 2000; Al-Ma-mary et al 2002) reported that the composition of beehiveproducts and its antioxidant capacity depend on several fac-tors, such as the flower source of the nectar, season, environ-mental factors as soil type and climate, genetic factors andfinally processing methods So, the possible health-relatedeffects due to the antioxidant activity of propolis may welldepend on its origin (Baltrusaityte et al 2007)

The RNA/DNA ratio values of the sea bass were rable with those measured in other studies (Alami-Durante

compa-et al.2007; Kerambrun et al 2012) The RNA/DNA ratio,

an index of nutritional status and larval growth (Buckley

et al 1999; Gwak et al 2003), increased from the controlgroup to those fed with 1.25 and 2.5 g kg 1 CPP-supple-mented diet This is in accordance with growth perfor-mance results where growth and protein synthesis werethe most expressed in fish at the higher CPP dietary

Table 3 Results of the two-way ANCOVA for serum biochemical parameters of European sea bass Dicentrarchus labrax juveniles fed with

dif-ferent levels of crude propolis powder (CPP) and subjected to an acute cold stressor

TP (g L 1 ) TRIG (mmol L 1 )

CHOL (mmol L 1 ) GLU (mmol L 1 )

supplementation To our knowledge, there is no literature

available on the effect of propolis on RNA/DNA ratio in

fish to substantiate our findings However, similar results

were observed with other immunostimulants that promote

higher RNA/DNA ratio in superior growing fish such as

probiotics (Bandyopadhyay & Mohapatra 2009), baker’s

yeast (Tewary & Patra 2011) and vitamin C (Tewary &

Patra 2008)

Before and after acute low-temperature stress, dietary

CPP supplementation resulted in significant alterations in

the level of serum biochemical parameters among

treat-ments Reduction in triglycerides and AST activities in

die-tary CPP supplementation groups before stress indicates

propolis hepatoprotective role, namely AST is an

ubiqui-tous aminotransferase in the mitochondrion of fish and is

an important index for the diagnosis of hepatopancreas

function and damage (Smet & Blust 2001) as elevation of

AST activity may indicate leakage of enzymes across

dam-aged plasma membranes and/or increased synthesis of

enzymes by the liver (Yang & Chen 2003) The

homoeosta-sis of lipids is one of the principal liver functions, and

plasma triglycerides increase can be used as an indicator of

liver disease, commonly associated with high fat content of

commercial fish pellet diet (Coz-Rakovac et al 2005) This

is in line with the findings of Deng et al (2011) where

sup-plementation of 1 g kg 1 ethanolic extract of propolis

sig-nificantly decreased plasma AST level and plasma

triglycerides level of rainbow trout After the cold stress,

serum triglycerides content increased in all groups

indicat-ing elevated energy demands to cope with the processes of

restoration, as using triglycerides under cold stress

(Kypria-nou et al 2010)

ALP, a polyfunctional transphosphorylase of organic

esters with important role in the skeleton mineralization of

aquatic animals and membrane transport activities (Bernet

et al.2001), in this study was positively correlated with the

propolis concentration in diet, and it was not affected by

cold stress Diet and feeding-dependent variation have been

demonstrated for ALP (Lemaire et al 1991; Bernet et al

2001; Wagner & Congleton 2004) showing positive

correla-tion with growth (Wagner & Congleton 2004; Ferri et al

2011), which is in accordance with our results as groups with

highest CPP feed supplementation demonstrated superior

growth performance and nutritional status Enhancement of

ALP serum activity has also been observed in heavy metal

exposure studies indicating mostly hepatic tissue damage

(Atli & Canli 2007); however, considering other biomarker

values in our study, hepatoprotective role is rather

sup-ported Kashkooli et al (2011) found no effects of ethanolic

propolis extract feed inclusion even in concentration up to

9 g kg 1 on juvenile rainbow trout serum biochemicalparameters, including ALP, but no growth stimulation dur-ing this 8-week trial was observed either

Blood cortisol levels are widely used as an indicator ofstress condition (Wendelaar Bonga 1997) because of theextreme sensitivity of the hypothalamo–pituitary–interrenal(HPI) axis The increase in serum cortisol level can be seen asthe sensitive signal of fish stress (Mommsen et al 1999) Asthe level of cortisol increases, blood glucose level might alsoincrease (Barton 2002; Begg & Pankhurst 2004) It is knownthat cortisol can regulate glucose mobilization due to corti-sol-dependent enhancement of gluconeogenesis (Vijayan

et al.1997; Mommsen et al 1999), accelerating glucose duction, which can be used to meet the increasing demandsfor energy from fish under stress In this study, serum corti-sol and glucose levels of all groups tended to increase whenfish were under low-temperature stress, similar to observa-tions in juvenile sea bass under temperature variation stress(Varsamos et al 2006) and subadult sea bass under short-term crowding stress (Di Marco et al 2008)

pro-At present, the effects of propolis on serum cortisol andglucose contents of fish under low-temperature stress havenot been reported; only relevant report of broiler chicksunder heat stress (Seven et al 2009) suggests that supple-mented dose of 3 mg kg 1 diet might prevent oxidativestress in the broilers exposed to heat In this study,2.5 g kg 1 CPP-supplemented diet alleviated the adverseeffects of cold-temperature stress on juvenile sea bassthrough significant reduction in cortisol response The pro-tective role of CPP might be related to its antioxidant effectand the ability to control the peroxidation of unsaturatedfatty acids, preventing the production of cholesterol, whichshowed no significant alterations throughout the experiment,and subsequent formation of cortisol (Kitabchi 1967)

In conclusion, 2.5 g kg 1 CPP-supplemented diet canenhance resistance to low-temperature stress and promotethe growth of sea bass, as indicated by the increase in feedconversion efficiency and RNA/DNA ratio, showing poten-tial for use in aquaculture industry Within that scope,further investigation of propolis well-established immuno-stimulatory properties on D labrax disease resistance iswarranted as well as cost-benefit analysis of using propolis

as feed additive

The idea and propolis preparation for this research wasconceived in collaboration with director of Hedera d.o.o

dr.sc Sa
sa Radic The authors would like to thank dr.sc.

Jasminka
Stefulj at the Ruder Bo
skovic Institute for

valu-able assistance during the experimental period This

research was supported by the Croatian Ministry of

Sci-ence, Education and Sports as part of the research

pro-gramme: Research on the development potential of

mariculture in Croatia (project no 001-0010501-0560)

Abd-El-Rhman, A.M.M (2009) Antagonism of Aeromonas

hydro-phila by propolis and its effect on the performance of Nile

tila-pia, Oreochromis niloticus Fish Shellfish Immunol., 27, 454–459.

Alami-Durante, H., Olive, N & Rouel, M (2007) Early thermal

history significantly affects the seasonal hyperplastic process

occurring in the myotomal white muscle of Dicentrarchus labrax

juvenile Cell Tissue Res., 327, 553–570.

Al-Mamary, M., Al-Meeri, A & Al-Habori, M (2002)

Antioxi-dant activities and total phenolics of different types of honey.

Nutr Res., 22, 1041–1047.

Atli, G & Canli, M (2007) Enzymatic responses to metal

expo-sures in a freshwater fish Oreochromis niloticus Comp Biochem.

Physiol C., 145, 282–287.

Bae, J.Y., Park, G.H., Lee, J.Y., Okorie, O.E & Bai, S.C (2012)

Effects of dietary propolis supplementation on growth

performance, immune responses, disease resistance and body

composition of juvenile eel, Anguilla japonica Aquacult Int., 20,

513 –523.

Baltrusaityte, V., Rimantas-Venskutonis, P & Ceksteryte, V.

(2007) Radical scavenging activity of different floral origin honey

and beebread phenolic extracts Food Chem., 101, 502–514.

Bandyopadhyay, P & Mohapatra, P.K.D (2009) Effect of a

pro-biotic bacterium Bacillus circulans PB7 in the formulated diets:

on growth, nutritional quality and immunity of Catla catla

(Ham.) Fish Physiol Biochem., 35, 467–478.

Banskota, A.H., Nagaoka, T., Sumioka, L.Y., Tezuka, Y., Awale,

S., Midorikawa, K., Matsushige, K & Kadota, S (2002)

Anti-proliferative activity of the Netherlands propolis and its active

principles in cancer cell lines J Ethnopharmacol., 80, 67–73.

Barton, B.A (2002) Stress in fishes: a diversity of responses with

particular reference to changes in circulating corticosteroids.

Integ Comp Biol., 42, 517–525.

Bazo, A.P., Rodrigues, M.A.M., Sforcin, J.M., Camargo, J.L.V.,

Ribeiro, L.R & Salvadori, D.M.F (2002) Protective action of

propolis on the rat colon carcinogenesis Teratog Carcinog

Mu-tag., 22, 183–194.

Begg, K & Pankhurst, N.W (2004) Endocrine and metabolic

responses to stress in a laboratory population of the tropical

damselfish Acanthochromis polyacanthus J Fish Biol., 64, 133–

145.

Berdalet, E., Rold an, C., Olivar, M.P & Lysnes, K (2005a)

Quan-tifying RNA and DNA in planktonic organisms with SYBR

Green II and nucleases Part A Optimisation of the assay Sci.

Mar., 69, 1–16.

Berdalet, E., Rold an, C & Olivar, M.P (2005b) Quantifying RNA

and DNA in planktonic organisms with SYBR Green II and

nucleases Part B Quantification in natural samples Sci Mar.,

69, 17–30.

Bernet, D., Schmidt, H., Wahli, T & Burkhardt-Holm, P (2001)

Effluent from a sewage treatment works causes changes in serum

chemistry of brown trout (Salmo trutta L.) Ecotoxicol Environ.

Saf., 48, 140–147.

Bricknell, I & Dalmo, R.A (2005) The use of immunostimulants

in fish larval aquaculture Fish Shellfish Immunol., 19, 457–472.

Buckley, L., Caldarone, E & Ong, T.L (1999) RNA –DNA ratio and other nucleic acid-based indicators for growth and condition

of marine fishes Hydrobiologia, 401, 265–277.

Caldarone, E.M., Clemmesen, C.M., Berdalet, E., Miller, T.J., Folkvord, A., Holt, G.J., Olivar, M.P & Suthers, I.M (2006) Intercalibration of four spectrofluorometric protocols for mea- suring RNA/DNA ratios in larval and juvenile fish Limnol Oce- anogr Methods, 4, 153–163.

Caruso, G., Denaro, M.G., Caruso, R., Mancari, F., Genovese, L &

Maricchiolo, G (2011) Response to short term starvation of growth, haematological, biochemical and non-specific immune parameters in European sea bass (Dicentrarchus labrax) and black- spot sea bream (Pagellus bogaraveo) Mar Environ Res., 72, 46–52.

C ß etin, E., Silici, S., Cßetin, N & G€ucßl€u, B.K (2010) Effects of diets containing different concentrations of propolis on hematological and immunological variables in laying hens Poult Sci., 89, 1703–

1708.

Chen, P.C & Wong, G (1996) Honey bee propolis: prospects in medicine Bee World, 77, 8–15.

Chen, L., Mehta, A., Berenvaum, M., Zangerl, A.R & Engeseth, J.

(2000) Honeys from different floral sources as inhibitors of matic browning in fruit and vegetable homogenates J Agric Food Chem., 48, 4997–5000.

enzy-Chu, W.H (2006) Adjuvant effect of propolis on immunization by inactivated Aeromonas hydrophila in carp (Carassius auratus gibelio) Fish Shellfish Immunol., 21, 113–117.

Cowan, M.M (1999) Plant products as antimicrobial agents Clin.

Microbiol Rev., 12, 564–582.

Coz-Rakovac, R., Strunjak-Perovic1, I., Hacmanjek, M., Popovic, N., Lipej, Z & Sostaric, B (2005) Blood Chemistry and Histo- logical Properties of Wild and Cultured Sea Bass (Dicentrarchus labrax) in the North Adriatic Sea Vet Res Commun., 29, 677–

687.

Cuesta, A., Rodriguez, A., Esteban, M  A & Meseguer, J (2005) In vivo effects of propolis, a honeybee product, on gilthead seabream innate immune responses Fish Shellfish Immunol., 18, 71–80.

De Castro, S.L (2001) Propolis: biological and pharmacological activities Therapeutic uses of this bee-product Ann Rev Biol.

Sci., 3, 49–83.

Deng, J., An, Q., Bi, B., Wang, Q., Kong, L., Tao, L & Zhang,

X (2011) Effect of ethanolic extract of propolis on growth formance and plasma biochemical parameters of rainbow trout (Oncorhynchus mykiss) Fish Physiol Biochem., 37, 959–967.

per-Denli, M., Cankaya, S., Silici, S., Okan, F & Uluocak, A.N.

(2005) Effect of dietary addition of Turkish propolis on the growth performance, carcass characteristics and serum variables

of quail (Coturnix coturnix Japonica) Asian-Aust J Anim Sci.,

18, 848–854.

Di Marco, P., Priori, A., Finoia, M.G., Massari, A., Mandich, A.

& Marino, G (2008) Physiological responses of European sea bass Dicentrarchus labrax to different stocking densities and acute stress challenge Aquaculture, 275, 319–328.

D €ugenci, S.K., Arda, N & Candan, A (2003) Some medicinal plants

as immunostimulant for fish J Ethnopharmacol., 88, 99–106.

Ferri, J., Topi c Popovic, N.,
Co
z-Rakovac, R., et al (2011) The effect of artificial feed on blood biochemistry profile and liver histology of wild saddled bream, Oblada melanura (Sparidae).

Mar Environ Res., 71, 218–224.

.