Aquaculture research, tập 42, số 12, 2011

Similarly, Gjen, Re-fstie, Ulla and Gjerde 1997 reported a high genetic correlation between resistance to furunculosis in a challenge test and survival during a ¢eld outbreak in Atlantic

Genetic variation in resistance to infectious pancreatic

a challenge test

Marte Wetten1, Sissel Kjglum1, Kjersti Turid Fjalestad1, Olaf Skjrvik2& Arne Storset1

1 Aqua Gen AS, Pir-Senteret,Trondheim, Norway

2 VESO Vikan,Vikan, Namsos, Norway

Correspondence: Sissel Kjglum, Aqua Gen AS, PO Box 1240, Pir-Senteret, N-7462 Trondheim, Norway E-mail: sissel.kjoglum@ aquagen.no

Abstract

In this study, we wanted to evaluate genetic variation

in resistance to infectious pancreatic necrosis (IPN)

in a breeding population of rainbow trout Two

hun-dred families were challenged at a commercial test

station, using methods that were previously

devel-oped for testing resistance to IPN in Atlantic salmon

Thirty-¢ve days after the challenge the accumulated

mortality was 26% The results show that resistance

to IPN is moderately heritable in the tested

popula-tion (h250.30) The genetic correlation between IPN

resistance and body weight was found to be low and

non-signi¢cant The signi¢cant additive genetic

var-iation found in IPN resistance after a controlled

chal-lenge test gives promise for successful breeding for

increased resistance to IPN in rainbow trout

Keywords: genetics, IPN, rainbow trout, variance

components

Introduction

Infectious pancreatic necrosis (IPN) is a serious

dis-ease in farmed salmonids and may cause major

eco-nomic losses both in fry and post-smolts In general,

IPN a¡ects rainbow trout (Oncorhynchus mykiss)

mainly in freshwater; only a few cases of IPN in

sea-running rainbow trout have been reported (Torkjel

Bruheim, pers comm.) This is di¡erent from Atlantic

salmon (Salmo salar) which is a¡ected both in

fresh-and seawater (Roberts & Pearson 2005) Although

good management and vaccination can reduce

mor-tality, the risk of disease outbreaks cannot be

elimi-nated Genetically improved resistance is therefore avaluable strategy to reduce losses due to IPN.Okamoto, Tayama, Kawanobe, Fujiki, Yasuda andSano (1993) described how progeny, whose parentshad experienced an IPN outbreak, had an increased re-sistance to IPN in rainbow trout By carefully checkingfor virus persistence in their breeding population, theywere able to establish an IPN-resistant strain with novirus in the ¢fth generation For commercial breeding,using survivors of either a natural outbreak or a con-trolled challenge is regarded as unacceptable due tothe risk of introducing IPN virus (IPNV) into the rear-ing facilities of the hatcheries For genetic improvementpurposes, selection based on survival information fromfull- and half-sibs has so far been the preferred methodfor commercial breeding organizations

In Atlantic salmon, genetic variation in resistance

to IPN has been well documented in di¡erent strains(Guy, Bishop, Brotherstone, Hamilton, Roberts,McAndrew & Woolliams 2006; Storset, Strand, Wet-ten, Kjglum & Ramstad 2007; Wetten, Aasmund-stad, Kjglum & Storset 2007; Guy, Bishop,Woolliams & Brotherstone 2009) In the Aqua Genpopulation of Atlantic salmon, selection for in-creased resistance to IPN has been very successfuland has led to a considerable genetic improvement

in IPN resistance (Storset et al 2007)

Both persisting losses due to IPN in commercialrainbow trout aquaculture and successful resultsfrom selective breeding for increased IPN resistance

in Atlantic salmon encouraged us to hypothesize thatthe challenge test developed for Atlantic salmonwould also enhance selective breeding of rainbowtrout The results in this study present the estimates

of variance components and an estimate of

heritabil-ity, and possibilities of genetic improvement of IPN

re-sistance by selection are discussed Also, because of

the importance of growth in the breeding programme

for rainbow trout, the genetic correlation between

body weight and IPN resistance was estimated

Materials and methods

Fish

The family-based breeding programme that Aqua

Gen is now managing was initiated in 1972 Fish were

collected from farms in both Norway and Sweden

during the years 1972^1974 to give the basis for three

distinct breeding populations of rainbow trout, one

for each year class in the generation interval (3 years)

In 2006, the three distinct breeding populations

were merged into one and another commercial strain

(previously bred separately) was also introduced to

the Aqua Gen breeding population

None of the four populations that have now been

merged into one have ever been selected for increased

IPN resistance

The 200 tested families in the 2006 year class were

made from112 dams and108 sires The mating design

was partial factorial, where eggs from each dam were

split in two and fertilized with milt from two di¡erent

males Due to few o¡spring in 24 of the families, only

200 out of the 224 families originally made could be

used for the challenge test Each family were

allo-cated to a separate compartment of an egg incubator

As all ¢sh challenged to IPNV were destroyed after

the test, individual body weight was recorded on

full-sibs of the challenged ¢sh 7 months after

fertiliza-tion In total,18952 ¢sh were registered for their body

weight Fish who had their body weights recorded

were kept as separate families in separate tanks from

fertilization until tagging by a PIT-tag (Trac id

sys-tems AS, Stavanger, Norway), at weight of 5^10 g

After tagging, all ¢sh were reared communally This

body weight registration is a routine in the breeding

programme, and the body weights registered at this

time are used in the breeding-value estimation

Challenge test

The ¢sh were challenged at VESO Vikan (Nord

Trn-delag, Norway), which is a large-scale research

la-boratory licensed to carry out challenge trials with

¢sh pathogens Before ¢rst feeding, un-vaccinated

fry from each family were transported as separate

groups to VESO Vikan The ¢sh had then a weight of

0.15^0.20 g The 200 families were challenged in parate tanks of10 L, and the number of ¢sh from eachfamily varied from 75 to 137 In total, mortalities fromIPN were registered on 22182 ¢sh

se-A pre-challenge of rainbow trout fry was performed

to see if they were susceptible to the IPNV isolate

V-1244 (results not shown) The inoculum originatedfrom a clinical case of IPN in Atlantic salmon post-smolts in Nord-Trndelag in 2001, and virus was pro-pagated to su⁄cient quantity using standard cell cul-ture methods at the Norwegian School of VeterinarySciences, Oslo, Norway The strain carries putativegenomic virulence marker motifs (Santi, Vakharia &Evensen 2004), and has been veri¢ed as highly viru-lent in Atlantic salmon fry challenge experiments Thisisolate has been used in challenge tests with Atlanticsalmon both in fry and post-smolts (Ramstad, Rom-stad, Knappskog & Midtlyng 2007; Storset et al 2007).The isolate was re-isolated from rainbow trout gonad(RTG-2) cells after passage through smolts by The Nor-wegian School of Veterinary Science and propagated asdescribed by Song, Santi, Evensen andVakharia (2005).The challenge test started after ¢rst feeding and 5days of acclimatization Fish were challenged using astandard bath challenge (Taksdal, Stangeland & Dan-nevig 1997) with a titre of 1.4 105TCID50mL 1 Attime of challenge the water £ow was stopped in alltanks The water was aerated and the water level waslowered to approximately 1.4 L The IPNV was added toproduce the concentrations given above Normal water

£ow was resumed after 3 h of bath challenge At mal £ow, the water £ow was minimum 0.7 Ltank 1min 1 The fry were fed manually twice a day(ad libitum) The challenge test lasted 35 days Watertemperature was 12 1C The level of oxygen in the in-coming water was logged Mortality was registered on

nor-a dnor-aily bnor-asis nor-and nor-all ¢sh thnor-at were not snor-ampled for i¢cation of IPN diagnosis were destroyed immediately

ver-To study the repeatability of the test, three of thefamilies were challenged in four replicates The repli-cates were randomly placed in the test facility.Samples from 40 randomly chosen tanks were ta-ken for veri¢cation of IPN diagnosis by using IPNV

Ag ELISATMTestline (TEST-LINE, Clinical tics, Brno, Czech Republic) (RodaØk, Posp|¤sil,TomaØnek,Vesely, Obr & Vajicek 1988)

Diagnos-Statistical analysesAlthough IPN resistance was de¢ned as a binarytrait, dead or alive, the trait is assumed to be poly-

genic This underlying continuous variable is often

called the liability trait (Falconer & Mackay 1996)

and is the justi¢cation for the use of a linear model

Several studies on statistical analyses of challenge

test data have been reported (Gitterle, degrd,

Gjerde, Rye & Salte 2006; degrd, Olesen, Gjerde &

Klemetsdal 2006, 2007; Kettunen, Serenius &

Fjales-tad 2007) Models considering test period survival

(both linear and threshold models), test-day survival

(linear repeatability) and time until death taking

cen-soring into account (Cox and Weibull proportional

hazard frailty) have been evaluated The general

con-clusions from these studies were that all models have

a high correlation between rankings of families A

linear model, as used in this paper, will therefore also

give the best families the highest breeding values

Also, the use of a linear model would simplify the

es-timation of the genetic correlation between body

weight and IPN resistance

The variance components were estimated with

REML method using the average information algorithm

by DMU package (Madsen & Jensen 2008) A bivariant

linear animal model was ¢tted Following model was

used for both IPN resistance and body weight:

Yij¼ mean þ tankjþ animaliþ residualij

where Yijis the observation on animal i within full-sib

group j; tankjis the random e¡ect of full-sib group j;

an-imaliis the random additive genetic e¡ect of the animal

i; residualijis the random residual for ¢sh i

The animal e¡ects were assumed to be  N(0,AsA2),

the common full-sib e¡ect were assumed to be

 N(0,IsT2) and the residuals were assumed to be

 N(0,IsE) The tank e¡ect includes the e¡ect of

common environment for full-sibs due to separaterearing of full-sib families before tagging, maternal ef-fects and non-additive e¡ects common to full-sibs ForIPN the tank e¡ect also include the e¡ect of tank in thechallenge test

The heritability (h2) for resistance to IPN andweight was estimated as

Eis an mate of the residual variance

esti-Heritability of IPN was adjusted to the underlyingliability scale (hu2) using the method of Robertson andLerner (1949):

h2¼ h2½pð1  pÞ=z2where hu2is the heritability on the underlying liabilityscale, p is the proportion of a¡ected ¢sh, z is theheight of the standard normal curve at the thresholdpoint

The tank e¡ect was estimated as the proportion ofvariance due to the tank e¡ect

ResultsChallenge testsThe accumulated mortality curve for all ¢sh in the frychallenge test is given in Fig 1 As can be seen fromthe graph, the daily mortalities are constant afteraround 9 days after infection At the termination ofthe test the overall mortality was 26.3%

0510

Figure 2 provides the mortality in each of the 200

families in the challenge test, sorted by mortality As

can be seen from the ¢gure, there was a large

varia-tion in mortality between the di¡erent families, with

the within family mortality ranging from 0% to 87%

The horizontal line in the ¢gure gives the overall

mortality in the test

Repeatability of the challenge test

Table 1 provides the mean mortality for the three

families in the challenge test that were replicated

four times each The results demonstrate a good

repeatability of the test

Body weight

Mean body weight at the time of registration was

61.7 g, with a standard deviation of 16.2 g Note that

individual body weight was registered on full-sibs of

the ¢sh in the IPN challenge test

Genetic analyses

Table 2 provides the variance components estimated

from the statistical analyses The heritability for IPN

re-sistance was estimated to be moderate and the

herit-ability of body weights was high When transformed

to the underlying liability scale, the heritability for

IPN resistance increased to 0.55 The variance due to

the common environment (tank e¡ect) was small for

both traits (Table 2) The genetic correlation between

IPN resistance and body weight was 0.20 (  0.10)

DiscussionThe results presented here demonstrate that a challengetest for IPN resistance developed for Atlantic salmonalso can be used for the breeding population of rainbowtrout investigated in this study IPN resistance in rain-bow trout was found to be a moderately heritable trait,and the trait has therefore potential for genetic improve-

0102030405060708090100

1 11 21 31 41 51 61 71 81 91 101 111 121 131 141 151 161 171 181 191

Family (sorting number)

Overall mortality

Figure 2 The distribution of mortalities in the families in the infectious pancreatic necrosis challenge test

Table 1 Mortalities (%) in the three replicated families, each family with four replicates

Parameters

IPN Body weight Estimate SE Estimate SE

Genetic variance (s A2) 0.06 0.01 118.26 18.17 Tank variance (s T2) 0.01 0.00 8.58 3.03 Residual variance (s E2) 0.13 0.01 134.70 9.30 Heritability observed scale 0.30 0.45 Heritability underlying scale 0.55

Proportion of variance caused by tank

ment by selection The genetic correlation between IPN

resistance and body weight of sib groups recorded seven

months after fertilization was found to be low

From the genetic analyses it was found that

around 30% of the variance in IPN resistance was

due to genetic e¡ects From studies in Atlantic

sal-mon, the heritabilities for IPN resistance have been

reported to be 0.31 using a linear animal model based

on challenge data (Wetten et al 2007) and from 0.07

to 0.56 using a reduced animal model based on ¢eld

data (Guy et al 2009) This is the ¢rst reported study

of genetic variation in IPN resistance after a

con-trolled challenge test in rainbow trout

One challenge of using controlled testing

for breeding purposes is the prediction of how the

results could improve survival in the ¢eld

Correla-tions in mortalities between experimental tests and

natural ¢eld outbreaks have not yet been published

in rainbow trout In Atlantic salmon, Wetten et al

(2007) described how the results from commercial

tests were validated by natural ¢eld outbreaks both

in ¢ngerlings and post-smolts They reported genetic

correlations between IPNV challenge test data and

¢eld mortality data in the range of 0.78^0.83 The

high genetic correlation implies that an IPNV

chal-lenge test on fry serves as a good predictor for IPN

¢eld mortality during both the juvenile stage and in

post-smolts in Atlantic salmon Similarly, Gjen,

Re-fstie, Ulla and Gjerde (1997) reported a high genetic

correlation between resistance to furunculosis in a

challenge test and survival during a ¢eld outbreak

in Atlantic salmon Further, Storset et al (2007)

de-scribed how the same ranking of low resistant and

high resistant groups of Atlantic salmon was

demon-strated in a fry test and a smolt test when full-sibs

were challenged to IPNV Their report also veri¢es a

positive e¡ect of selection based on challenge-tests

in the fry-stage on resistance at later life-stages

In Atlantic salmon, the virulence of IPNV serotype

Sp strains can be classi¢ed as virulent, moderately

virulent, or low virulent, depending on two major

outer capsid protein (VP2) residues  217 and 221

(Song et al 2005) Previous challenge experiments in

Atlantic salmon have demonstrated that the most

IPN susceptible Aqua Gen families are a¡ected by

virulent strains only, whereas moderately virulent

strains give very low mortality in ¢sh from the Aqua

Gen breeding population (Santi et al 2004; Song et al

2005) The IPNV isolate used in the challenge test of

this study has been classi¢ed as virulent in Atlantic

salmon Results from ¢eld outbreaks, however,

indi-cate that strains non-virulent for Atlantic salmon

could be highly virulent for rainbow trout (Nina

San-ti, pers comm.) A further development of the IPNVchallenge test for rainbow trout could therefore be

to challenge ¢sh by other strains of IPNV However,the survival of the fry in a challenge study of suchduration would mainly depend upon the activation

of the innate immune system (Dorson, De Kinkelin

& Torchy 1992; Tatner 1996; Ellis 2001), and it is thusexpected that the genetic resistance should be e¡ec-tive against a range of di¡erent IPNV isolates.The potential to improve resistance to IPN also de-pends on the genetic correlations between IPN andother traits in the breeding programme In the pre-sent study, the genetic correlation between IPN resis-tance and weight at the smolt stage was found to besmall and negative but with a large standard error.The standard error of the estimate could be largerthan expected since body weights and IPN resistancewere not registered on the same ¢sh, but on full-sibs.However, in the breeding programme, breeding va-lues for the selection candidates are estimated based

on their own record of body weight, and records onmortalities from full-sibs So for breeding purposes,this correlation is relevant From the literature, ge-netic correlations between growth and disease resis-tance have been reported to be both zero andunfavourable Leeds, Silverstein,Vallejo, Palti, Rexro-

da III, Evenhuis, Hadidi, Weber, Welch and Wiens(2009) studied response to selection on bacterial coldwater disease resistance in rainbow trout and found

no signi¢cant genetic correlation between the ease and body weights However, Henryon, Jokum-sen, Berg, Lund, Pedersen, Olesen and Slierendrecht(2002) reported unfavourable correlations ( 0.14 to

dis- 0.33) between the predicted breeding values for sistance to viral haemorrhagic septicaemia in rain-bow trout and the predicted breeding values forbody weight, body length and feed conversion e⁄-ciency

re-A challenge in breeding for genetic resistance todiseases is that the trait is only scored as two cate-gories, dead or alive However, if the date of death isalso registered, the number of statistical methods forevaluating such data is increased For breeding pur-poses, the correct ranking of the families is impera-tive as well as the possibility to estimate geneticcorrelations between the traits of importance Sev-eral papers evaluating statistical methods for ana-lyses of binary data conclude that the ranking offamilies are the same when di¡erent models includ-ing linear models are used (Gitterle et al 2006; de-grd et al 2006)

A successful breeding programme for increased

resistance may not only reduce the number of

in-fected ¢sh at any time, but at the same time also

re-duce the risk of susceptible ¢sh being infected The

number of susceptible individuals in the population

plays an important role in determining the level of

herd or £ock immunity, which in turn determines

whether an organism can survive in the population

Flock immunity can be demonstrated when just a

portion of a population is vaccinated and when this

portion provides protection to unvaccinated animals

as well (Martin, Meek & Willeberg 1987)

There are reasons to believe that selection for

in-creased IPN resistance also will reduce the number

of healthy carriers in the population Aqua Gen has

in their brood ¢sh found a much lower frequency of

IPNV carriers in brood ¢sh selected for increased

re-sistance to IPN than in brood ¢sh not selected for this

trait (Nina Santi, pers comm.) If the number of

healthy carriers is reduced, this could possibly reduce

the infection pressure on both farmed and wild ¢sh

Breeding from survivors of a commercial challenge

is considered unacceptable, because of the risk of

in-troducing IPNV into hatcheries, and also because of

the chance of using infected carriers as brood ¢sh

However, by challenging full-sibs of the brood ¢sh, it

is possible to do selection on brood ¢sh that remain

unchallenged This method has already been proven

to be very successful in Atlantic salmon (Wetten et al

2007) Although a selection programme based on

challenge testing and possibly selection on several

other traits simultaneously does not allow for rapid

development of resistant strains, the accumulated

im-proved resistance over generations can be substantial

References

Dorson M., De Kinkelin P & Torchy C (1992) Interferon

synthesis in rainbow trout fry following infection with

infectious pancreatic necrosis virus Fish and Shell¢sh

Im-munology 2, 311^313.

Ellis A.E (2001) Innate host defence mechanisms of ¢sh

against viruses and bacteria Developmental and

Compara-tive Immunology 25, 827^839.

Falconer D.S & Mackay T.F.C (1996) Introduction to

Quantita-tive Genetics, 4th edn Longman Group, Essex, UK.

Gitterle T., degrd J., Gjerde B., Rye M & Salte R (2006)

Ge-netic parameters and accuracy of selection for resistance

to White Spot Syndrome Virus (WSSV) in Penaeus

(Litope-naeus) vannamei using di¡erent statistical models

Aqua-culture 251, 210^218.

Gjen H.M., Refstie T., Ulla O & Gjerde B (1997) Genetic relations between survival of Atlantic salmon in chal- lenge and ¢eld tests Aquaculture 158, 277^288.

cor-Guy D.R., Bishop S.C., Brotherstone S., Hamilton A., Roberts R.J., McAndrew B.J & Woolliams J.A (2006) Analysis of the incidence of infectious pancreatic necrosis mortality

in pedigreed Atlantic salmon, Salmo salar L, populations Journal of Fish Diseases 29, 637^647.

Guy D.R., Bishop S.C., Woolliams J.A & Brotherstone S (2009) Genetic parameters for resistance to infectious pancreatic necrosis in pedigreed Atlantic salmon (Salmo salar) post-smolts using a reduced animal model Aqua- culture 290, 229^235.

Henryon M., Jokumsen A., Berg P., Lund I., Pedersen P.B., Olesen N.J & Slierendrecht W.J (2002) Genetic variation for growth rate, feed conversion e⁄ciency, and disease re- sistance exists within a farmed population of rainbow trout Aquaculture 209, 59^76.

Kettunen A., Serenius T & Fjalestad K.T (2007) Three tical methods for genetic analysis of disease resistance against vibriosis in Atlantic cod (Gadus morhua L) Journal

statis-of Animal Science 85, 305^313.

Leeds T.D., Silverstein J.T., Vallejo R.L., Palti Y., Rexroda C.E III, Evenhuis J., Hadidi S., Weber G.M., Welch T.J & Wiens G.D (2009) Response to selection for bacterial cold water dis- ease resistance in rainbow trout (abstract) The 10th Interna- tional Symposium on Genetics in Aquaculture 22^26 June 2009,Thailand.

Madsen P & Jensen J (2008) A User’s Guide to DMU.Version 6, Release 4.7 Faculty of Agricultural Science, Department

of Genetics and Biotechnology, University of Aarhus, hus, Denmark.

Aar-Martin S.W., Meek A.H & Willeberg P (1987) Descriptive demiology In: Veterinary Epidemiology (ed by J.-P Vaillancourt), pp 79–120 Iowa State University Press, Ames, IA, USA.

epi-degrd J., Olesen I., Gjerde B & Klemetsdal G (2006) luation of statistical models for genetic analysis of chal- lenge test data on furunculosis resistance in Atlantic salmon (Salmo salar): prediction of ¢eld survival Aquacul- ture 266,70^76.

degrd J., Olesen I., Gjerde B & Klemetsdal G (2007) luation of statistical models for genetic analysis of chal- lenge-test data on ISA resistance in Atlantic salmon (Salmo salar): prediction of progeny survival Aquaculture 266,70^76.

Eva-Okamoto N., Tayama T., Kawanobe M., Fujiki N., Yasuda Y & Sano N (1993) Resistance of a rainbow trout strain to in- fectious pancreatic necrosis Aquaculture 117,71^76 Ramstad A., Romstad A.B., Knappskog D.H & Midtlyng P.J (2007) Field validation of experimental challenge models for IPN vaccines Journal of Fish Diseases 30, 723^731.

Robertson A & Lerner I.M (1949) The heritability of all-or-none traits; viability of poultry Genetics 34, 395^411.

Roberts R.J & Pearson M.D (2005) Infectious pancreatic

ne-crosis in Atlantic salmon, Salmo salar L Journal of Fish

Diseases 28, 383^390.

RodaØk L., Posp|¤sil Z.,TomaØnek J.,Vesely T., Obr T & Vajicek L.

(1988) Enzyme-linked immunosorbent assay (ELISA) for

the detection of infectious pancreatic necrosis virus

(IPNV) in culture £uids and tissue homogenates of the

rainbow trout, Salmo gairdneri Richardson Journal of Fish

Diseases 11, 225^235.

Santi N.,Vakharia V N & Evensen  (2004) Identi¢cation of

putative motifs involved in the virulence of infectious

pancreatic necrosis virus.Virology 322, 31^40.

Song H., Santi N., Evensen O & VakhariaV.N (2005)

Molecu-lar determinants of infectious pancreatic necrosis virus

virulence and cell culture adaptation Journal of Virology

79, 10289^10299.

Storset A., Strand C., Wetten M., Kjglum S & Ramstad A.

(2007) Response to selection for resistance against

infec-tious pancreatic necrosis in Atlantic salmon (Salmo salar, L.) Aquaculture 272, 62^68.

Taksdal T., Stangeland K & Dannevig B.H (1997) Induction

of infectious pancreatic necrosis (IPN) in Atlantic salmon Salmo salar and brook trout Salvelinus fontinalis by bath challenge of fry with infectious pancreatic necrosis virus (IPNV) serotype Sp Diseases of Aquacultured Organisms

28, 39^44.

Tatner M.F (1996) Natural changes in the immune system of

¢sh In: Fish Physiology,Vol.15 (ed by S Hoar, D.J Randall

& A.P Farrell), pp 255^287 Academic Press, San Diego,

CA, USA.

Wetten M., Aasmundstad T., Kjglum S & Storset A (2007) Genetic analysis of resistance to infectious pancreatic ne- crosis in Atlantic salmon (Salmo salar L.) Aquaculture

272, 111^117.

Effects of low dietary protein level on serum oestradiol, testosterone and sex reversal in rice field eel,

Hanwen Yuan1, Shiyuan Gong1, Zhangjie Chu2, Guobin Zhang1,Yongchao Yuan1, Wenjie Gong1&Jianlin Yan1

1 College of Fisheries, Key Lab of Agricultural Animal Genetics, Breeding and Reproduction of Ministry of Education, Huazhong Agricultural University,Wuhan, Hubei, China

2 Zhejiang Ocean University, Zhoushan, Zhejiang, China

Correspondence: Shiyuan Gong, College of Fisheries, Huazhong Agricultural University,Wuhan, Hubei Province 430070, China E-mail: gsy@mail.hzau.edu.cn

Abstract

We investigated the e¡ects of low dietary protein in

isocaloric diets on sex reversal of Monopterus albus

by evaluating the oestradiol (E2) and testosterone (T)

concentrations, gonadosomatic index (GSI), sex ratio

and gonad structure at the histological level Fish

(9.50 1.50 g average initial weight; n 5 3 per group)

were fed with ¢ve practical diets containing 100, 150,

200, 250 or 400 g kg 1crude protein to apparent

sa-tiation for 15 months Serum E2and T concentrations

were determined by radioimmunoassays E2

concen-trations and GSI signi¢cantly increased while T

con-centrations decreased as the dietary protein level was

raised Fish fed 400 g kg 1of dietary protein had

signi¢cantly higher E2concentrations and GSI than

those fed with lower dietary protein levels The

T concentrations of ¢sh fed 100 g kg 1of dietary

pro-tein was signi¢cantly higher than that of ¢sh fed

higher dietary protein levels The shift of sex ratio

to-wards more male and intersex ¢sh was observed with

decreasing dietary protein levels Therefore, low

diet-ary protein level may promote sex change from

fe-male to fe-male in M albus This study provides

important information for successful reproductive

management and may be exploited for aquaculture

of this species

Keywords: Monopterus albus, low dietary protein

level, oestradiol, testosterone, gonadosomatic

in-dex, sex reversal

IntroductionThe rice ¢eld eel, Monopterus albus, taxonomicallybelongs to the teleosts, the family Synbranchidae ofthe order Synbranchiformes (Neoteleostei, Teleostei,Vertebrata), and is also the only representative spe-cies of the group of Synbranchidae in China It inha-bits mainly China, Japan and Southeast Asia (Li, Liao,

Yu, Cheng & Tong 2007), but it can also be found inNorthern Australia and Southeastern United States(Collins, Trexler, Nico & Rawlings 2002) This freshwater ¢sh is an economically important species inSoutheast Asia Owing to its unique characteristics,such as relative small genome size and natural sex re-versal from female via intersex into male during itslife span, it is a good model for studying vertebrates

in comparative genomics, evolution and tal biology, especially in sexual development (Zhou,Cheng & Tiersch 2002)

developmen-The rice ¢eld eel is a protogynous hermaphrodite,strictly changing its sex unidirectionally from func-tional female to male naturally during development.Moreover, the intersex gonads are found during itsgrowth for an extended period of time (Liem 1963;Chan & Philips1967; Liem1968) The breeding season

is from May to August The rice ¢eld eels spend 1 ormore years reaching puberty under natural ambientconditions, which are all females in the beginning.After spawning they normally begin to change sexfrom female to intersex and then ¢nally to male Theeels may not develop into mature functional males

until they are more than 3 years old In fact, females

of more than 6 years old have been found in the

nat-ural population (Xiao 1995; Yang, Chen, Ruan & Su

2008) These special characteristics have made the

rice ¢eld eel an ideal species for the study of sex

deter-mination and di¡erentiation (Zhou, Cheng, Zhang,

Guo, Cooper & Tiersch 2002) Since the discovery of

natural sex reversal in rice ¢eld eels in 1944 (Liu

1944), some e¡orts have been made to uncover the

mechanism underlying the process in this species

from both the physiological and biochemical

perspec-tive (Liu & Ku 1951; Chan,Wai,Tang & Lofts 1972; Liu,

Wan, Su, Zhang & Han 1987; Yeung & Chan 1987; Liu,

Cui,Wang & Chen 1990; Tao, Lin, Kraak & Peter 1993;

Yeung, Chen & Chan 1993a, b; Fan, Cai, Lin & Zhang

1999; Zou 2000), at the cytological level (Xiao 1993,

1995; Xiao & Liu 1995) and at the molecular level

(Lu, Cheng, Guo & Zhou 2003; Wang, Cheng, Xia,

Guo, Huang & Zhou 2003; Yu, Cheng, Guo, Xia &

Zhou 2003; Zhou, Liu, Guo, Yu, Cheng, Huang,

Tiersch & Berta 2003; Xia, Cheng, Yu, Guo & Zhou

2004; Huang, Guo, Shui, Gao, Yu, Cheng & Zhou

2005; Jang, Zhou, Xia, Zhao, Cheng & Zhou 2006)

However, the detailed mechanism of sex reversal of

this species remains unclear

Because of over¢shing, environmental destruction

or other unknown factors, the resources of rice ¢eld

eels have drastically decreased over the past few

dec-ades in China (He, Liu, Guo, Jin & Zhang 2004;Yin, Li,

Zhou & Liu 2005), arti¢cial propagation and breeding

is an e¡ective means for supplementing the natural

resources To date, the farming scale of this species is

limited by scarcity of the eel fry since large-scale

arti-¢cial propagation and breeding is hindered by the

shortage of mature male broodstock and low

abso-lute fecundity Therefore, studies designed to

under-stand the process of sex reversal in ¢sh with low

absolute fecundity will be important role for the

breeding industry Some studies have been

con-ducted on genetic variation and population

di¡eren-tiation of the species (Liu, Wang, Zeng, Luo & Han

2005; Yang, Zhou, Zhang & Li 2005; Li et al 2007)

Liem (1963) presumed that adverse environmental

factors such as intermittent drought and food

short-age may promote sex change of rice ¢eld eel Shi,

Lin & Tang (1998) reported that long-term (6 weeks)

starvation improved the sex change process in the

species

Thus far, little is known about the e¡ect of low

pro-tein levels on sex reversal of rice ¢eld eel In the

pre-sent study, we propose that a low dietary protein level

promotes sex change from female to male in

protogy-nous rice ¢eld eel The objectives of our study were toevaluate the serum oestradiol (E2), testosterone (T)concentrations and sex reversal of M albus fed with

a low protein diet This study should provide usefulinformation to obtain su⁄cient numbers of suitablemales for successful reproductive managementand eventual increase of the species in the farmingindustry

Materials and methodsDiet preparation

Five experimental diets containing 100, 150, 200, 250

or 400 g crude protein per kg were formulated(Table 1).White ¢shmeal and soybean meal were used

as protein sources, ¢sh oil as a lipid source and

Table 1 Ingredients used in the preparation of tal diets and their estimated nutrient and energy contents (400 g kg 1protein level was used as the control diet)

experimen-Dietary protein levels (g kg 1)

100 150 200 250 400

Ingredients (g kg  1 ) American white fishmeal 20 100 180 260 500 Soybean meal 120 120 120 120 120

unless otherwise speci¢ed): vitamin A,

65 000 IU; vitamin D 3 , 45 000 IU; vitamin E, 25; vitamin K 3 , 5; vitamin B 1 , 12.5; vitamin B 2 , 12.5; vitamin B 6 , 15; vitamin B 12 , 0.025; DL-calcium pantothenate, 40; niacin, 50; folic acid, 2.5; bio- tin, 0.08; inositol, 75; ascorbic acid, 120.

zContains (as g kg  1

in diet): Ca, 100; P, 50; K, 30; Na, 20; Mg, 10;

Fe, 22; Zn, 3; Mn, 3; Cu, 1.8; Co, 0.15; I, 0.12; Se, 0.05.

‰Mean values from three replicates.

zComputed as 16.7 kJ g  1 protein and carbohydrate and 37.6 kJ g  1 lipid.

kProtein to energy ratio in mg kJ  1 CMC, carboxymethyl cellulose.

a-starch as a carbohydrate source The 400 g kg 1

protein level feed was used as control diet (Yang

2002) The diets contained energy contents estimated

at values of 16.7, 37.6 and 16.7 kJ g 1for proteins,

lipids and carbohydrates respectively (Garling &

Wilson 1976) Fish oil was gradually added to all dry

ingredients which were mixed in a commercial food

mixer for 30 min Approximately 400 mL of water

per kg of feed was slowly blended into the mixture

which was further processed through a laboratory

strip-extruder to 2.0 mm diameter in size The feed

was dried overnight at room temperature and stored

at 20 1C until used

Feeding and sampling

The experiment was carried out in net cages which

were ¢xed in one earthen pond Hatchery-reared rice

¢eld eel (M albus) ¢ngerlings were collected from the

Shayang New Century Fisheries Technology

(Shayang City, Hubei Province, China), reared in net

cages (6 m 2 m  1.5 m) for 3 weeks, and

gradu-ally weaned from earthworms onto the formulated

feed After the pre-acclimation period, a total of 750

¢sh with average initial body weights of 9.50 1.50 g

from the same spawning were randomly placed into

15 net cages (2 m 1m  1.5 m) at 50 ¢sh per cage,

and acclimated to the formulated feed for 2 weeks

Three cages of ¢sh were fed with each of the ¢ve feed

treatments described above Alternanthera

philoxer-oides (an aquatic plant) were cultured in each net

cage to imitate the natural habitat

During the experiment, the ¢sh were hand-fed once

daily at 18:00 hours until apparent satiation except on

days with thunderstorm or water temperatures in

ex-cess of 10^38 1C Each cage was fed with the

formu-lated diet until no ¢sh came for the food and the

leftovers of the food was removed daily The ¢sh were

cultured for 15 months, from May 2008 to July 2009

Three ¢sh were sampled from each cage randomly at

monthly intervals from the fourth month

Water quality parameters were monitored in the

net cages weekly Temperatures ranged from 4 to

36.8 1C, pH from 6.2 to 7.8, ammonia nitrogen was

lower than 0.03 mg L 1and dissolved oxygen was

noto5 mg L 1

Blood sampling

Blood samples were drawn from the caudal vein

using a disposable syringe equipped with a gauge

needle The volume of the collected blood rangedfrom 500 to 3000mL with a mean volume of

1200 200 mL The collected blood samples wereimmediately transferred to 1.5 mL disposable centri-fuge tubes Both syringes and centrifuge tubes wererinsed with 10 mg mL 1heparin in 0.9% NaCl Bloodsamples were centrifuged at 1200 g for 25 min, andthe serum was frozen and stored at 20 1C for radio-immunoassays

Determination of gonadosomatic index (GSI)Each ¢sh was weighed ( 0.01g) before blood wascollected The gonads were then removed andweighed ( 0.0001g) The GSI was calculated ac-cording to the formula: GSI (%) 5 Gonad weight (g)/

¢sh total weight (g) 100

Sex ratio and survival ratesSex ratio and survival rates were calculated at theend of the experiment Sexing was conducted by bothmacroscopical and microscopical methods

Measurement of E2and T

E2and T were assessed by following the protocols ofthe radioimmunoassay kits provided by BeijingNorth Institute of Biological Technology (Beijing,China) The respective hormones were labelled with125

I and then used as radioactive markers The ples were run in the same assay Measurements of in-traassay and interassay precision (coe⁄cients ofvariation) wereo10% and 15% for E2and T respec-tively The sensitivities of the T and E2assays wereo2 pg mL 1 All of these analyses were the same asthose used in our recent study (Chu, Gong, Zhang,Zhang,Yuan & Yuan 2009)

sam-Gonad histologyAfter the gonads were weighed, tissue samples were

¢xed in Bouin’s ¢xative for 36 h, dehydrated and bedded in para⁄n wax Thick transverse sections(7mm) were stained with Ehrlich’s haematoxylin andeosin (H&E) and examined under a light microscope(Kokokiris, Brusle, Kentouri & Fostier 1999) The go-nads were further classi¢ed on the basis of the stage

em-of the germ cell

Statistical analysis

Analyses of variance for hormone concentrations

and GSI were performed usingPROC MIXED(SAS

Insti-tute 1999) The di¡erences in the mean sex ratios and

survival rates between treatment groups were

as-sessed by one-way analysis of variance (ANOVA)

fol-lowed by Duncan’s multiple range test at 0.05 level

(Duncan 1955)

Results

E¡ects of di¡erent protein levels on E2and T

concentrations

Both E2and T concentrations had the same trend in

the reproductive cycle of the rice ¢eld eels (Figs 1 and

2), where E2and T concentrations had a decreasing

tendency in each group after spawning; moreover,

the rates of the decreases were fast at ¢rst and then

slowed thereafter E2 concentrations achieved the

lowest levels in January, while T concentrations could

not be detected in December and January in any

group E2 and T concentrations increased quickly

after the overwintering period, achieved the peak in

May and June and then decreased

The values of E2and T concentrations and GSI in

Table 2 were estimated as means from the 12 months

during the experimental period Di¡erent dietary

protein levels signi¢cantly a¡ected both E2and

Tcon-centrations of the rice ¢eld eels fed with the

formu-lated feeds E2concentrations of the control group

were signi¢cantly higher than the other groups

There were no signi¢cant di¡erences for E2

concen-trations of ¢sh fed with the 200, 150 and 100 g kg 1

protein levels T concentrations of the control groupwere signi¢cantly lower than the other treatmentgroups

T concentrations were signi¢cantly higher in ¢shreceiving the 100 g kg 1protein level diet than thosewith the 200, 250 and 400 g kg 1 dietary treat-ments However, there was no signi¢cant di¡erencefor T concentrations between 250 and 200 g kg 1dietary treatments In the experiment, E2concentra-tions achieved minimum and maximal values89.56 15.20 and1240.22  323.86 pg mL 1, whichoccurred in the 100 g kg 1protein level group inJanuary and 400 g kg 1protein level group in Mayrespectively T concentration achieved the maximalvalue of 804.02 34.84 pg mL 1in the 100 g kg 1protein level group in June

To analyse the e¡ect of E2and T concentrations indi¡erent months, we conducted the variance analysis

on the dietary protein levels months interaction(Table 3) Signi¢cant di¡erences between E2and Tconcentrations were detected at di¡erent dietary pro-tein levels The dietary protein levels, month e¡ectand the dietary protein levels months interactione¡ects proved to be signi¢cant

E¡ects of di¡erent protein levels in diets onGSI

The GSIs of di¡erent groups had the same trend in productive cycle of the rice ¢eld eels (Fig 3) The GSIshad a decreasing tendency in each group afterspawning Before spawning, GSIs began to increase,which were not coincident in di¡erent protein levels.The GSIs were a¡ected signi¢cantly by the formu-lated feeds in this experiment (Table 2) The highest

re-Figure 1 E¡ects of di¡erent dietary protein levels on

ser-um E2 concentrations in Monopterus albus in di¡erent

months The spawning season is from May to August Data

are means of three replicate cages

Figure 2 E¡ects of di¡erent dietary protein levels on

ser-um T concentrations in Monopterus albus in di¡erentmonths Data are means of three replicate cages

GSIs were observed in ¢sh fed with the 400 g kg 1

protein diet No signi¢cant di¡erences were found in

GSIs among the 250, 200 and150 g kg 1dietary

pro-tein levels.While GSIs of ¢sh fed with the 250 g kg 1

protein diet was signi¢cantly higher compared with

those with the 100 g kg 1 protein diet The GSIs

achieved the minimum and maximal values,

0.611 0.35% and 11.669  3.66%, which occurred

in the 100 g kg 1dietary protein level group in ary and 400 g kg 1dietary protein level group in Julyrespectively The GSIs of the ¢sh fed with higher diet-ary protein levels increased faster before and duringthe spawning period in the experiment

Janu-To analyse the e¡ect of GSIs in di¡erent months, weconducted the variance analysis on the dietary pro-tein levels months interaction (Table 3) Signi¢-cant di¡erences among GSIs were detected atdi¡erent dietary protein levels.We found that the diet-ary protein levels, month e¡ect and the dietary pro-tein levels months interaction e¡ects were allsigni¢cant

E¡ects of di¡erent protein levels in diets onsex ratio and survival rates

Sex ratio and survival rates of the rice ¢eld eels fedwith di¡erent dietary protein levels varied (Table 4).Signi¢cant di¡erences were observed in the femaleratio at di¡erent dietary protein levels, the highest fe-male ratio occurred at the 400 g kg 1dietary proteinlevel, followed by 250, 200, 150 and 100 g kg 1diet-

Table 2 Variance analysis for E 2 and T concentrations and gonadosomatic index (GSI) with di¡erent dietary protein levels Protein levels (g kg 1) E 2 (pg ml 1)  T (pg ml 1

Values within the column are signi¢cantly di¡erent at the 0.05 probability level.

wValues within the column are signi¢cantly di¡erent at the 0.01 probability level.

Figure 3 E¡ects of di¡erent dietary protein levels on

go-nadosomatic index (GSI) of Monopterus albus in di¡erent

months Data are means of three replicate cages

ary protein levels The intersex ratio of those fed with

the 400 and 250 g kg 1dietary protein levels were

lower than those of the other groups Male ratios of

the 400 g kg 1dietary protein level group (5.34

4.64%) were signi¢cantly lower than the other

groups The highest male ratio obtained at the

100 g kg 1 dietary protein level was signi¢cantly

di¡erent compared with the 250 g kg 1protein level

Meanwhile, the survival rates of the 100 g kg 1

diet-ary protein level were signi¢cantly lower than those

of the other groups

E¡ects of di¡erent protein levels in diets on

gonad development

Histological structure of the female gonads from all

treatment groups showed the same pattern during

the experimental period (Fig 4) After the spawning

period, mature oocytes began to degenerate (Fig 4a)

and produced some empty spaces (Fig 4b) in ovaries

During the overwintering period, many females had

ovaries with many primary and secondary yolk stage

oocytes and empty spaces (Fig 4c) Before and during

spawning period, the female gonads were ¢lled at the

vitellogenic stage and fully grown oocytes stage (Fig

4d^f) After the overwintering period, the intersex

and male gonads were found in the experimental ¢sh

(Fig 5) In the intersex ¢sh, some spermatocytes

could be observed among the numerous oocytes in a

few gonads from the lower dietary protein level

groups (Fig 5a).With the development of the gonads,

the testicular tissue expanded and the ovarian tissue

gradually shrunk (Fig.5b) The mature testis were

ob-served in lower dietary protein level treatment

groups (Fig 5c), and fortunately we observed the

spawning of the fully functional males in the

150 g kg 1dietary protein level treatment group in

the following July (Fig 5d)

The ovarian tissue occupied the whole gonad ing the spawning season After spawning season, theovarian tissue began to regress And the bisexual go-nad which showed some spermatocytes in the middle

dur-of numerous oocytes was ¢rst found in the

200 g kg 1protein level group in March 2009 Thetesticular tissue expanded to become the dominanttissue in the bisexual gonad in the100 g kg 1proteinlevel group in May The complete sex change of thetestis was ¢rst observed in the 200 g kg 1protein le-vel group in June and some males had well-developedtesticular tissue which indicated that the sex reversal

in the ¢sh resulted in functional males

Discussion

In the present study, serum E2and T concentrations

of the rice ¢eld eels had the same trend during the productive cycle After the spawning period, E2and Tconcentrations began to decrease and achieved thelowest point during the overwintering period, whilethey increased before the spawning period andreached peak levels during the spawning period Si-milar results were reported for rice ¢eld eel M albus

re-by Song and Xiong (1993) In our study, serum E2and

T concentrations were in£uenced by di¡erent dietaryprotein levels E2concentration increased as the diet-ary protein level was raised, while T concentrationwas inversely correlated with dietary protein level.According toYamamoto (1969), sex steroids are thenatural inducers of gonadal di¡erentiation in tele-osts, which require changes in steroidogenesis, pre-ceding or coinciding with gonadal di¡erentiation It

is widely accepted that sex steroid hormones play animportant and a speci¢c role during the process ofsex di¡erentiation in ¢sh Many studies on sex ster-oids regulation of gametogenesis of teleost ¢sh have

Table 4 Results of variance analysis for sex ratio and survival with di¡erent dietary protein levels

Sex ratio (%) Protein levels (g kg 1) Female Intersex Male Survival ratio (%)

identi¢ed that oestrogen is produced by the ovary to

regulate oocyte growth, while androgens are

pro-duced by the testis to regulate spermatogenesis

(Na-gahama 1994, 1997) Chang, Lee and Chen (1994)

investigated the seasonal changes of steroid pro¢les,

E2, T and 17a-hydroxyprogesterone in black porgy,

Acanthopagrus schlegeli, and suggested that high

le-vels of plasma E2in the prespawning and spawning

season were likely correlated with the natural sex

change of black porgy E2likely plays an important

role in sex reversal of protandrous black porgy

As T is the hormonal precursor of

11-ketotestoster-one (a key sex steroid closely associated with

testicu-lar function in ¢sh) and E2, its role in sexual

di¡erentiation is therefore of great importance and isgenerally re£ected in the processes of sex di¡erentia-tion (Chang, Hung, Chiang & Lan 1999) Lau, Lin,Lee, Sun, Dufour and Chang (1997) and colleaguesreported that an increase of testosterone engendered

by testosterone administration resulted in testiculargrowth and male function in protandrous blackporgy E2is considered as a major sex steroid in fe-male ¢sh and plays an important role in vertebrateprenatal development, particularly concerning sex-ual di¡erentiation and sex change (Chang, Lau & Lin1995; Chang & Lin 1998; Piferrer 2001; Lee,Yueh, Du,Sun & Chang 2002; Lange, Hartel & Meyer 2003; Wu,Tomy, Nakamura & Chang 2008) E2has been re-

Figure 4 Transverse sections of gonads of adult female Monopterus albus (a) Ovary of 250 g kg 1protein level treated

¢sh in September, showing degenerating oocytes (arrow) (scale bar 51200mm) (b) Ovary of 150 g kg 1protein level ted ¢sh in October, showing the degenerated oocytes producing empty spaces (arrow) (scale bar 5 800mm) (c) Ovary of

trea-200 g kg 1protein level treated ¢sh in December (scale bar 5 800mm) (d) Ovary of 400 g kg 1protein level treated ¢sh

in February (scale bar 5 800mm) (e) Ovary of 250 g kg 1protein level treated ¢sh in April (scale bar 51200mm) (f)Ovary of 400 g kg 1protein level treated ¢sh in June (scale bar 51200mm)

ported to control physiological processes in the

go-nads and to induce vitellogenin synthesis in liver

(Blaise, Gagne¤, Pellerin & Hansen 1999; Gagne¤,

Blaise, Pellerin, Pelletier, Douville, Gauthier-Clerc &

Viglino 2003)

The present study demonstrated that with di¡erent

dietary protein levels, the concentration of E2

in-creased, while the concentration of Tdecreased Both

of them increased quickly before and during the

spawning period These results indicated that the

hormones could play a role as endogenous

modula-tors of gametogenesis These transient increases at

the levels of sex steroids during gametogenesis

re-£ected the function of steroids as endogenous

modu-lators of reproduction Bogart (1987) suggested that

gonadal sex was determined by a local ratio of

andro-gen to oestroandro-gen with relatively higher or lower ratios

during testicular or ovarian development

respec-tively Sex di¡erentiation was shown to be closely

as-sociated with this ratio in some species (Zheng,Wang,

Zhao, Zuo & Chen 2005; Rougeot, Krim, Mandiki,

Kestemont & Me¤lard 2007)

As indicated by our results above, GSI increased

with the increasing dietary protein level, with the

in-£uence being more apparent in females than inmales The GSI was in£uenced signi¢cantly by diet-ary protein levels before and during the spawningperiod As described previously, we found that serumsteroids were in£uenced by protein levels

In the present study, sex ratios after various dietaryprotein level treatments were determined by histolo-gical examination of the gonads Intersex and maleratios were higher in lower dietary protein levelgroups than the control group at the end of the ex-periment The results indicated that higher proteinlevels inhibited the sex di¡erentiation in rice ¢eld eeland promoted the maturation process in females.The highest proportion of males was found in the

100 g kg 1dietary protein level group Sex nation in certain ¢sh may be in£uenced by a singlevariable, while multiple stimuli have a combinatoriale¡ect in other species (Nakamura, Kobayashi, Chang

determi-& Nagahama 1998; Baroiller, Guiguen determi-& Fostier 1999;Pandian & Koteeswaran 1999) The sex ratio isa¡ected by sexual di¡erentiation Temperature, ster-oid hormones and social cues are some exogenousfactors that could also in£uence the sex ratio in a

¢sh population (Fishelson 1970; Yamamoto 1999;

Figure 5 Transverse sections of gonads of intersex and male Monopterus albus (a) Intersex gonads from ¢sh treated with

200 g kg 1protein level in March, showing spermatocytes (arrow) in the middle of the numerous oocytes (scalebar 5100mm) (b) Intersex gonads from ¢sh treated with 100 g kg 1protein level in May, showing spermatocytes (arrow)

in the middle of the numerous oocytes (scale bar 5100mm) (c) Testis containing spermatocytes (arrow) from ¢sh treatedwith 250 g kg 1protein level in June (scale bar 5100mm) (d) Testis of 150 g kg 1protein level treated ¢sh (after spawn-ing) in July (scale bar 5100mm)

Van, Verheyen & Witters 2003) Santiago, Aldaba

and Laron (1982) reported that survival ratio was

in£uenced by dietary protein levels in Tilapia nilotica

fry In our study, ¢sh fed with lower protein diets

had higher mortality rates, which could be due to

nutritional imbalance and a poor food conversion

ratio

Gonad size and the number of oocytes decreased

dramatically in females after spawning Thereafter,

the oocytes began to shrink and produced some

empty spaces between the oocytes in the ovaries A

few rudimentary oocytes and empty spaces could be

observed in the gonads during the overwintering

period The histological structure of the female

go-nads in this study showed that the gogo-nads changed

with di¡erent dietary protein levels We found that

the oocytes degenerated faster, and the number of

oocytes was less in lower protein level groups than

that in the control group Korfsmeier (1969) reported

follicular epithelial cells could take part in the

phago-cytosis and degradation of degenerating oocytes in

zebra¢sh In the present study, we found that the

go-nad size and the number of oocytes were larger in the

control group than that in the lower protein level

groups; however, there was no apparent di¡erence of

the egg size After the overwintering period, the

number of oocytes increased gradually, and the

empty spaces were ¢lled with oocytes again The

oo-cytes then proliferated quickly, and the ovaries

reached the peak size The observations of the

devel-opment of oocytes in the reproductive cycle in our

study were similar to those reported previously (Liu

et al 1990; Xiao & Liu 1995; Yang, Gu, Wang & Dong

2004)

In the present study, the bisexual gonad which

showed some spermatocytes in the middle of

numer-ous oocytes was ¢rst found in the 200 g kg 1protein

level group in March 2009 We observed the ovarian

tissue regressed, and the testicular tissue expanded

to become the dominant tissue in the bisexual gonad

in the 100 g kg 1protein level group in May The

complete sex reversal of the testis was ¢rst observed

in the 200 g kg 1protein level group in June and

some males had well-developed testicular tissue,

in-cluding spermatozoa Additionally, we also observed

testis during postspawning (July to September)

peri-ods which indicated that the sex reversal in the ¢sh

resulted in functional males Histological analysis of

the gonads showed a few spermatocytes in the

mid-dle of the numerous oocytes, which indicated that

the gonad was in a transitional phase.With the

devel-opment of intersexual gonads, spermatogenic germ

cells proliferation occurs progressively, and eration of ovarian tissue proceeds gradually The de-veloping spermatocytes and oocytes can exist in thegonads at the same time The development of testicu-lar tissue in the gonads during the transitional phase

degen-is progressive and invariably combined with ovariantissue at various stages of degeneration Finally, thetestis replaces the ovary in the gonads, indicatingthat the change of the gonads from female to male iscomplete The results of the present study suggestthat the ¢sh were functional females for the ¢rst re-productive season During the postspawning andnonspawning periods after the ¢rst reproductive sea-son, the ovarian tissue regressed, and the testiculartissue expanded to become the dominant tissue inthe bisexual gonad Ultimately, the ¢sh had well-de-veloped testicular tissue and no ovarian tissue in thegonad The timing of onset of sex change was impos-sible to detect by observing the ¢sh The completion

of sex reversal in each ¢sh was con¢rmed by nation of the gonad Previous researchers studied theinteraction between ovarian and testicular tissueand sex change in protandrous black porgy Theyfound that the E2-stimulated early sex change ¢shhad a reversible sex change after E2termination andsuggested that the testicular tissue e¡ected on the de-velopment of ovarian tissue in the bisexual gonad(Chang & Yueh 1990; Lee, Huang & Chang 2008;

exami-Wu & Chang 2009) Uchida, Yamashita, Kitanoand Iguchi (2002) suggested the mechanism of testi-cular and ovarian di¡erentiation in zebra¢sh to beinduced by oocyte apoptosis However, the mechan-ism of sex reversal in rice ¢eld eel has not been welldocumented

In conclusion, this investigation has demonstratedthat a low dietary protein level could signi¢cantly af-fect E2and T concentration, GSIs, sex ratio and gonaddevelopment under our experimental conditions Theresults indicate that a low dietary protein level couldpromote sex change from female to male This ¢ndingsupports the hypothesis that food shortage may pro-mote sex change of M albus (Liem 1963) We believethe information obtained in our study will contribute

to the e¡orts to increase reproductive management,help facilitate aquaculture of the species and even-tually increase the productivity to meet the demand

of the market The study also provides new evidencefor producing male rice ¢eld eels However, furtherinvestigations will be required to fully reveal the me-chanisms of sex reversal of rice ¢eld eel in order to de-sign rational methods for controlling reproduction ofthis species in aquaculture

This study was supported by the Key Technologies R

& D Program during the 11th Five-Year Plan (No

2006AA2003A01).We thank ZhaojiaYuan for his

as-sistance with ¢sh culture and sampling, Dr Liwu

Zhang for his assistance with the data analysis, and

Prof Yanxiu Liu for helpful comments on the

manu-script

References

Baroiller J.F., Guiguen Y & Fostier A (1999) Endocrine and

environmental aspects of sex di¡erentiation in ¢sh

Cellu-lar and MolecuCellu-lar Life Sciences 55, 910^931.

Blaise C., Gagne¤ F., Pellerin J & Hansen P.D (1999)

Determi-nation of vitellogenin-like properties in Mya arenaria

he-molymph (Saguenay Fjord, Canada): a potential

biomarker for endocrine disruption Environmental

Toxi-cology 14, 455^465.

Bogart M.H (1987) Sex determination: a hypothesis based

on steroid ratios Journal of Theoretical Biology 128, 349^

357.

Chan S.T.H & Philips J.G (1967) The structure of the gonad

during natural sex reversal in Monopterus albus (Pisces:

Teleostei) Journal of Zoology 151, 129^141.

Chan S.T.H.,Wai S.O.,Tang F & Lofts B (1972) Biopsy studies

on the nature sex reversal in Monopterus albus

(Pisces:Tel-eostei) Journal of Zoology 167, 415^421.

Chang C.F & Yueh W.S (1990) Annual cycle of gonad

histol-ogy and steroid pro¢les in the juvenile males and adult

fe-males of the protandrous black porgy, Acanthopagrus

schlegeli Aquaculture 91, 179^196.

Chang C.F & Lin B.Y (1998) Estradiol-17b stimulates

aroma-tase activity and reversible sex change in protandrous

black porgy, Acanthopagrus schlegeli The Journal of

Experi-mental Zoology 280, 165^173.

Chang C.F., Lee M.F & Chen G.R (1994) Estradiol-17b

asso-ciated with the sex reversal in protandrous black porgy,

Acanthopagrus schlegeli The Journal of Experimental

Zool-ogy 268, 53^58.

Chang C.F., Lau E.L & Lin B.Y (1995) Stimulation of

sperma-togenesis or of sex reversal according to the dose of

exo-genous estradiol-17b in juvenile males of protandrous

black porgy, Acanthopagrus schlegeli General and

Compara-tive Endocrinology 100, 355^367.

Chang C.F., Hung C.Y., Chiang M.C & Lan S.C (1999) The

concentrations of plasma sex steroids and gonadal

aro-matase during controlled sex di¡erentiation in grey

mul-let, Mugil cephalus Aquaculture 177, 37^45.

Chu Z.J., Gong S.Y., Zhang G.B., Zhang L., Yuan Y.C & Yuan

Z.J (2009) E¡ects of estradiol valerate on steroid

hor-mones and sex reversal of female rice ¢eld eel, Monopterus

albus Journal of theWorld Aquaculture Society (accepted).

Collins T.M., Trexler J.C., Nico L.G & Rawlings T.A (2002) Genetic diversity in a morphologically conservative inva- sive taxon: multiple introductions of swamp eels to the Southeastern United States Conservation Biology 16, 1024^1035.

Duncan D.B (1955) Multiple-range and multiple F tests metrics 11, 1^42.

Bio-Fan J.Y., Cai H.F., Lin Y.H & Zhang X.X (1999) The serum protein contents and electrophoretic analysis at di¡erent developing stages of Monopterus albus Journal of Moun- tain Agriculture and Biology 18, 216^218 (in Chinese) Fishelson L (1970) Protogynous sex reversal in the ¢sh An- thias squamipinnis (Teleostei, Anthiidae) regulated by the presence or absence of a male ¢sh Nature 227, 90 Gagne¤ F., Blaise C., Pellerin J., Pelletier E., Douville M., Gau- thier-Clerc S & Viglino L (2003) Sex alteration in soft- shell clams (Mya arenaria) in an intertidal zone of the Saint-Lawrence River (Que¤bec, Canada) Comparative Bio- chemistry and Physiology Part C: Comparative Pharmacol- ogy and Toxicology 134, 189^198.

Garling D.L & Wilson R.P (1976) Optimum dietary protein

to energy ratios for channel cat¢sh ¢ngerlings, Ictalurus punctatus The Journal of Nutrition 106, 1368^1375.

He S.L., Liu X.W., Guo Z.L., Jin H & Zhang J.P (2004) On the genetic diversity of three species of Monopterus Journal of Hunan Agricultural University 30, 145^147 (in Chinese) Huang X., Guo Y.Q., Shui Y., Gao S., Yu H.S., Cheng H.H & Zhou R.J (2005) Multiple alternative splicing and di¡er- ential expression of dmrt1 during gonad transformation

of the rice ¢eld eel Biology of Reproduction 73, 1017^1024 Jang S.H., Zhou F., Xia L.X., Zhao W., Cheng H.H & Zhou R.J (2006) Construction of a BAC library and identi¢cation of Dmrt1 gene of the rice ¢eld eel, Monopterus albus Bio- chemical and Biophysical Research Communications 348, 775^780.

Kokokiris L., Brusle S., Kentouri M & Fostier A (1999) ual maturity and hermaphroditism of the red porgy, Pa- grus pagrus (Teleostei, Sparidae) Marine Biology 134, 621^629.

Sex-Korfsmeier K.H (1969) Uber die degeneration von eizellen des zebra¢sches, Brachydanio rerio, in vitro Cell andTissue Research 98, 99^105.

Lange I.G., Hartel A & Meyer H.H.D (2003) Evolution of trogen functions in vertebrates The Journal of Steroid Bio- chemistry and Molecular Biology 83, 219^226.

oes-Lau E.L., Lin B.Y., Lee F.Y., Sun L.T., Dufour S & Chang C.F (1997) Stimulation of testicular function by exogenous testosterone in male protandrous black porgy, Acanthopa- grus schlegei Journal of Fish Biology 51, 327^333 Lee M.F., Huang J.D & Chang C.F (2008) The development of ovarian tissue and female germ cells in the protandrous black porgy ¢sh, Acanthopagrus schlegeli Zoological Stu- dies 47, 302^316.

Lee Y.H.,Yueh W.S., Du J.L., Sun L.T & Chang C.F (2002) omatase inhibitors block natural sex change and induce male function in the protandrous black porgy, Acanthopa-

Ar-grus schlegeli Bleeker: possible mechanism of natural sex

change Biology of Reproduction 66, 1749^1754.

Li W.T., Liao X.L.,Yu X.M., Cheng L & Tong J (2007) Isolation

and characterization of polymorphic microsatellites in a

sex-reversal ¢sh, rice ¢eld eel (Monopterus albus)

Molecu-lar Ecology Notes 7,705^707.

Liem K.F (1963) Sex reversal as a natural process in the

syn-branchiform ¢sh Monopterus albus Copeia 2, 303^312.

Liem K.F (1968) Geographical and taxonomic variation in

the pattern of natural sex reversal in the teleost ¢sh order

Synbranchiformes Journal of Zoology 156, 225^238.

Liu C.K (1944) Rudimentary hermaphroditism in the

sym-branchoid eel, Monopterus javanensis Sinensia 15, 1^8.

Liu C.K & Ku K.Y (1951) Histological changes in the gonad of

Monopterus javanensis during sex transformation

Sinen-sia 2, 85^109.

Liu L.G., Wang W.B., Zeng B.P., Luo Y.S & Han Q (2005)

RAPD analysis of Monopterus albus (Zuiew) populations

with three colors Fisheries Science 24, 22^25.

Liu R.Z.,Wan H., Su B.R., Zhang F.C & Han D.S (1987)

Rela-tionship between sex reversal and serum proteins in

Monopterus Acta Hydrobiologica Sinica 11, 22^28.

Liu X.Y., Cui T.C., Wang L.C & Chen J.C (1990) Histological

and ultrastructural changes in the gonad during sex

re-versal in Monopterus albus Acta Hydrobiologica Sinica 14,

166^171.

Lu H., Cheng H.H., GuoY.Q & Zhou R.J (2003) Two alleles of

the Sox9a2 in the rice ¢eld eel The Journal of Experimental

Zoology Part B: Molecular and Developmental Evolution

299, 36^40.

Nagahama Y (1994) Endocrine regulation of gametogenesis

in ¢sh International Journal of Developmental Biology 38,

217^229.

NagahamaY (1997) 17a, 20b-Dihydroxy-4-pregnen-3-one, a

maturation-inducing hormone in ¢sh oocytes:

mechan-isms of synthesis and action Steroids 62, 190^196.

Nakamura M., Kobayashi T., Chang X.T & Nagahama Y.

(1998) Gonadal sex di¡erentiation in teleost ¢sh Journal

of Experimental Zoology 281, 362^372.

Pandian T.J & Koteeswaran R (1999) Lability of sex

di¡eren-tiation in ¢sh Current Science 76, 580^583.

Piferrer F (2001) Endocrine sex control strategies for the

feminization of teleost ¢sh Aquaculture 197, 229^281.

Rougeot C., Krim A., Mandiki S.N.M., Kestemont P & Me¤lard

C (2007) Sex steroid dynamics during embryogenesis and

sexual di¡erentiation in Eurasian perch, Perca £uviatilis.

Theriogenology 67, 1046^1052.

Santiago C.B., Aldaba M.B & Laron M.A (1982) Dietary

crude protein requirement of Tilapia nilotica fry Philippine

Journal of Biology 11, 255^265.

Shi Q., Lin H.R & Tang P.L (1998) E¡ects of starvation on

gonadal development and serum melatonin levels in the

rice eel, Monopterus albus Zuiew Journal of Beijing Normal

University (Natural Science) 34, 395^398 (in Chinese).

Song P & Xiong Q.M (1993) The annual cycle of serum

es-tradiol and testosterone in spontaneous sex-reversing ¢sh

Monopterus albus (Zuiew) Journal of Wuhan University (Natural Science) 2, 115^119 (in Chinese).

TaoY.X., Lin H.R., Kraak G.V.D & Peter R.E (1993) Hormonal induction of precocious sex reversal in the rice¢eld eel, Monopterus albus Aquaculture 118, 131^140.

Uchida D.,Yamashita M., Kitano T & Iguchi T (2002) Oocyte apoptosis during the transition from ovary-like tissue to testes during sex di¡erentiation of juvenile zebra¢sh Journal of Experimental Zoology 205,711^718.

Van D.B.K.,Verheyen R & Witters H (2003) E¡ects nylestradiol in a partial life-cycle test with zebra¢sh (Danio rerio): e¡ects on growth, gonads and female reproductive success Science of theTotal Environment 309, 127^137 Wang R., Cheng H.H., Xia L.X., Guo Y.Q., Huang X & Zhou R.J (2003) Molecular cloning and expression of Sox17 in gonads during sex reversal in the rice ¢eld eel, a teleost

of17a-ethy-¢sh with a characteristic of natural sex transformation Biochemical and Biophysical Research Communications

303, 452^457.

Wu G.C & Chang C.F (2009) wnt4 is associated with the velopment of ovarian tissue in the protandrous black porgy, Acanthopagrus schlegeli Biology of Reproduction 81, 1073^1082.

de-Wu G.C., Tomy S., Nakamura M & Chang C.F (2008) Dual roles of cyp19a1a in gonadal sex di¡erentiation and devel- opment in the protandrous black porgy, Acanthopagrus schlegeli Biology of Reproduction 79, 1111^1120 Xia L.X., Cheng H.H., Yu H.S., Guo Y.Q & Zhou R.J (2004) Molecular cloning and expression of the osteoclast-stimu- lating-factor-like gene from the rice ¢eld eel The Journal of Experimental Zoology Part B: Molecular and Developmental Evolution 302, 174^181.

Xiao Y.M (1993) Study on the reproductive biology of pterus albus (Zuiew) 1: early gonadogenesis and structure change Monopterus albus Journal of Natural Science of Hu- nan Normal University 16, 346^349 (in Chinese) Xiao Y.M (1995) Study on the reproductive biology of Mono- pterus albus (Zuiew) 2: female development of Monopterus albus Journal of Natural Science of Hunan Normal Univer- sity 18, 45^51 (in Chinese).

Mono-XiaoY.M & Liu Y (1995) Study on the histology in sex ging from intersex to male of Monopterus albus (Zuiew) Journal of Fisheries of China 19, 297^304 (in Chinese) Yamamoto E (1999) Studies on sex-manipulation and pro- duction of cloned populations in hirame, Paralichthys oli- vaceus (Temminck et Schlegel) Aquaculture 173, 235^246 Yamamoto T (1969) Sex di¡erentiation In: Fish Physiology (ed by W.S Hoar & D.J Randall), pp 117^175 Academic Press, NewYork, NY, USA.

chan-Yang D.Q (2002) Study on nutritional requirements and tive enzymes of Monopterus albus PhD thesis, Huazhong Agricultural University, China.

diges-Yang D.Q., Chen F., Ruan G.L & Su Y.B (2008) Relationship between sex reversal, body weight and age of Monopterus albus Journal ofYangtze University (Natural Science Edition) Agricultural Science 4, 45^47.

Yang F.Q., Zhou Q.B., ZhangY.P & Li X.H (2005) Variation of

esterase isoenzyme among three kinds of bodily color

Monopterus albus in Poyang Lake Region Journal of

Eco-nomic Animal 9, 110^113 (in Chinese).

Yang W.Y., Gu Z.Q., Wang C.H & Dong W.Z (2004)

Observa-tion of gonad histology in the course of the sex reverses

for the eels (Monopterus albus) Progress inVeterinary

Med-icine 25, 113^115 (in Chinese).

Yeung W.S.B & Chan S.T.H (1987) The plasma sex steroid

pro¢les in the freshwater, sex-reversing teleost ¢sh,

Mono-pterus albus (Zuiew) General and Comparative

Endocrinol-ogy 65, 233^242.

YeungW.S.B., Chen H & Chan S.T.H (1993b) E¡ect of LH and

LHRH-analog on gonadal development and in vitro

steroi-dogenesis in the protogynous Monopterus albus General

and Comparative Endocrinology 89, 323^332.

Yeung W.S.B., Chen H & Chan S.T.H (1993a) The in vitro

metabolism of radioactive androstenedione and

testo-sterone by the gonads of the protogynous Monopterus

albus at di¡erent sexual phases: a time-course and

seaso-nal study General and Comparative Endocrinology 89, 313^

322.

Yin S.W., Li J.Z., Zhou G.J & Liu Y (2005) Population genetic

structure of rice ¢eld eel (Monopterus albus) with RAPD

markers Chinese Journal of Applied Environmental Biology

11, 328^332 (in Chinese).

Yu H.S., Cheng H.H., Guo Y.Q., Xia L.X & Zhou R.J (2003) Alternative splicing and di¡erential expression of P450c17 (CYP17) in gonads during sex transformation in the rice ¢eld eel Biochemical and Biophysical Research Communications 307, 165^171.

Zheng R.H.,Wang C.G., ZhaoY., Zuo Z.H & Chen Y.X (2005) E¡ect of tributyltin, benzo(a)pyrene and their mixture ex- posure on the sex hormone levels in gonads of cuvier (Se- bastiscus marmoratus) Environmental Toxicology and Pharmacology 20, 361^367.

Zhou R.J., Cheng H.H., Zhang Q.Y., Guo Y.Q., Cooper R.K & Tiersch T.R (2002) SRY-related genes in the genome of rice ¢eld eel (Monopterus albus) Genetique Selection Evolu- tion 34, 129^137.

Zhou R.J., Cheng H.H & Tiersch T.R (2002) Di¡erential ome duplication and ¢sh diversity Reviews in Fish Biology and Fisheries 11, 331^337.

gen-Zhou R.J., Liu L., Guo Y.Q., Yu H.S., Cheng H.H., Huang X., Tiersch T.R & Berta P (2003) Similar gene structure of two Sox9a genes and their expression patterns during go- nadal di¡erentiation in a teleost ¢sh, rice ¢eld eel (Mono- pterus albus) Molecular Reproduction and Development 66, 211^217.

Zou J.X (2000) Analysis of the relationship between sexual reversal and serum proteins Reservoir Fisheries 1, 13^15 (in Chinese).

Stages of rock bream oplegnathus fasciatus

(Temminck et Schlegel 1844): embryonic development

Tao He1,2, Zhi Z Xiao1, Qing H Liu1, Dao Y Ma1, Shi H Xu1, Yong S Xiao1& Jun Li1

1

Center of Biotechnology R&D, Institute of Oceanology, Chinese Academy of Sciences, Qingdao, China

2 Graduate school, Chinese Academy of Sciences, Beijing, China

Correspondence: J Li, Center of Biotechnology R&D, Institute of Oceanology, Chinese Academy of Sciences, Qingdao 266071, China E-mail: junli@ms.qdio.ac.cn

Abstract

The embryonic development of rock bream,

Opleg-nathus fasciatus, was studied from fertilization until

hatching The hatching occurred approximately at

25 h after fertilization at 23.5 0.5 1C The

embryo-genesis is divided into seven stages: Zygote period,

Cleavage period, Morula period, Blastula period,

Gas-trula period, Segmentation period and Hatching

per-iod The ¢rst cleavage furrow of rock bream fertilized

eggs is vertically oriented, as is usual until horizontal

cleavage occurs at the ¢fth cleavage The blastocoel is

observed between the blastoderm and I-YSL at

blas-tula period At 90%-epiboly stage, the earliest somitic

furrow appears in the middle of embryo The

myo-tomes develop from somites at 15 h 30 min post

ferti-lization The Kup¡er’s vesicle consisting of ventrally

I-YSL and dorsally columnar cells appears with the

completion of epiboly It degenerates gradually with

the penetration of some eosinophilic granules and

disappears completely at 20 h 30 min after

fertiliza-tion The digestive tract, a straight tubule, is

di¡eren-tiated into foregut, midgut and hindgut The

epithelium of midgut and hindgut are the monolayer

cubic and columnar cells respectively The staging

series provides a preliminary baseline reference for

future studies on embryos of the rock bream

Keywords: oplegnathus fasciatus, embryo,

develop-mental stage, histology

Introduction

Histo-morphological and physiological studies on

onto-geny (embryos, larvae and juveniles) are imperative for

understanding basic developmental mechanisms of eost species and improving ¢ngerling-production proto-cols The development of zebra¢sh Danio rerio (Kimmel,Ballard, Kimmel, Ullmann & Schilling 1995) and meda-

tel-ka Oryzias latipes (Iwamatsu 2004) has been studied indetail as the ‘model’organisms The studies on develop-ment of other cultured species are crucial for providinguseful information on reproductive biology, larval diets,organogenesis and metamorphosis In contrast to‘mod-el’species, studies on the early development of many cul-tured species are limited, largely because it is thehistorically di⁄culty to obtain complete developmentalseries of marine ¢sh Recently, the rapid development ofaquaculture makes it possible to investigate the earlydevelopment of marine ¢sh systematically In recentyears, ontogenetic studies on striped trumpeter Latrislineatea (Furlani & Ruwald 1999), Japanese conger Con-ger myriaster (Horie, Utoh, Yamada, Okamura, Zhang,Mikawa, Akazawa, Tanaka & Poka 2002), summer

£ounder Paralichthys dentatus (Martinez & Bolker2003), sunbleak Leucaspius delineatus (Pinder & Gozlan2004), annual ¢sh Austrolebias viarius (Arezo, Pereiro &Berois 2005),striped snakehead Channa striatus (Mari-muthu & Hani¡a 2007) and Port Jackson shark Hetero-dontus portusjacksoni (Rodda & Seymour 2008) weremorphologically described In marine ¢sh aquaculture,the egg quality is the key to gain the production of high-quality larvae and juveniles (Franz & Pierpaolo 2004),and the embryonic development, the ¢rst step of fryrearing, is one of the most important parameters forevaluating egg quality However, the embryogenesis athistology level was rarely reported (Morrison, Miyake

& Wright 2001; Hall, Smith & Johnston 2004).The rock bream, Oplegnathus fasciatus, a subtropi-cal and carnivorous species, is an economically im-

portant marine ¢sh in East Asia It has recently been

targeted as a candidate species for commercial

aqua-culture and stock enhancement Especially in China,

the culture scale of rock bream has increased

gradu-ally to meet the market demand in the last decade

The high commercial and ornamental value make it

a promising aquaculture species in the future

How-ever, to some extent, the lack of information on

onto-genetic development has restricted the breeding

industry of this species

In recent years, many studies have been carried

out to investigate the nutrition (Wang, Kim, Bai,

Huh & Cho 2003; Nam, Cho, Choi, Kim, Kim & Kim

2005; Shan, Quan & Dou 2008; Lim & Lee 2009),

dis-ease (Yoshikoshi & Inoue 1990; Jung & Oh 2000; Lee,

Kim, Kim, Nam, Kim & Kim 2004; Cho, Choi, Kim,

Kim, Kim, Bang & Nam 2006; Choi, Kwon, Nam, Kim

& Kim 2006; Kim, Ha, Ahn, Nam, Kim & Kim 2007)

and morphological features (Chang, Mao, Wu &

Zhang 2005; Quan & Xiao 2007; Liu, Xu,Wang, Lu &

Qu 2008; Xiao, Zheng, Yu & Li 2008) of rock bream

embryo and larvae For rock bream, we have

success-fully domesticated the wild species caught from the

East Sea of China Manipulation of environmental

cues (such as photoperiod and water temperature)

and nutrient supply had been used to induce its

re-production, and the fertilized eggs were obtained In

previous studies, we have preliminarily investigated

the arti¢cial breeding technique and observed the

morphological features in the embryonic

develop-ment with a stereoscopic microscope However, the

histological traits in embryogenesis have not been

observed

In order to provide valuable information on

devel-opmental biology of marine ¢sh and establish a

sta-ging series for describing the accurate timing of

embryogenesis, we observed the embryonic

develop-ment of rock bream by histology in combination with

gross morphology in this paper

Material and methods

Fish and rearing conditions

Adults of rock bream (5 years old) were stocked in a

laboratory tank with aerated sea water and natural

light Water temperature was at 23 1C and

photoper-iod of 15L:9D induced gonadal maturation Fertilized

eggs of the same naturally spawned clutch were

col-lected by a bloting-silk net and incubated at

23.5 0.5 1C with weak aeration in natural seawater

(32 0.5%)

Morphological and histological observationEmbryonic development was recorded with a stereo-scopic microscope equipped with a NikonYS100 digi-tal still camera The embryogenic process of rockbream was divided into seven stages: Zygote period,Cleavage period, Morula period, Blastula period,Gastrula period, Segmentation period and Hatchingperiod

About 500 fertilized eggs every 30 min were ¢xed

in Bouin’s ¢xative overnight, and then transferred to70% ethanol for longtime preservation After beenoriented in agar, samples were dehydrated in agraded series of ethanol and embedded in para⁄n.About 4^6mm sagittal and cross sections were cut

by a KEDEE 1508A slicer, and then stained with matoxylin and eosin (H&E) basically, and Mallory’strichrome stain as a supplementary explanation forhistological observation

hae-ResultsFertilized oocytesThe fertilized egg of rock bream is transparent with ayellow yolk, a brown-yellow lipid droplet and a smallperivitelline space The diameters of the egg and oilglobule are 0.845 0.055 and 0.175  0.025 mm re-spectively The fertilized eggs are free £oating and theunfertilized ones sink down

Zygote period (0^30 min)

As the fertilization ¢nished, the cytoplasm coveringthe yolk begins to accumulate at the animal poleand forms a disc-shaped bulge, which is called theblastoderm It is clearly distinguishable from the yolkmass

Cleavage period (30 min^2 h 30 min)Two-cell stage (30 min)

As in other teleosts, cleavage in rock bream is blastic The ¢rst cleavage furrow is vertically or-iented, as is usual until the 32-cell stage The ¢rstmitosis divides the blastoderm into two blastomeres

mero-of equal size (Fig 1, 1) and they appear otherwise distinguished from each other The two centrosomes,and the spindle ¢bres that grow-out from the centro-somes (Fig 2, 1), are distinctly visible in an optic mi-croscope

Four-cell stage (50 min)

The second cleavage is also vertical and four

blasto-meres form in a 22 array with the cleavage furrow

at a right angle to the ¢rst cleavage plane The size ofthe four blastomeres is equal to one another Fromthe animal pole view, the blastoderm is ellipsoidal inshape (Fig 1, 2)

Eight-cell stage (1h 10 min)

The third cleavage involves two cleavage planes that

are parallel to the ¢rst one dividing the four

blasto-meres into a 2 4 array of asymmetrical cells (Fig

1, 3, 2, 2) The blastoderm is bilaterally symmetrical

and elongates along the axis of the second cleavage

plane

16 -cell stage (1h 30 min)

Just as the third one, the fourth cleavage includes two

cleavage planes, and they are parallel to the second

cleavage planes The two cleavage planes divide the

2 4 array of blastomeres into 4  4 array of cells

(Fig 1, 4), which consist of four central blastomeres

and 12 marginal ones

32-cell stage (1h 50 min)

At this stage, the ¢fth cleavage planes divide the fourcentral blastomeres horizontally into eight cells thatform two layers (an inner and an outer layer) and the

12 marginal blastomeres vertically into 24 cells, soonly 28 cells can be counted from the animal poleview (Fig 1, 5) This uneven cleavage makes the blas-toderm partially strati¢ed with a single layer on theperiphery and two cell layers in the centre From thisstage onwards, the mitoses of blastomeres becomeasynchronous

64 -cell stage (2 h)Because of the horizontal cleavages, at this stage thearray of blastomeres looks like the 32-cell stage on

5

YI-YSL

Bl

7

Ep

HyBc

M

En Bc

M

6a

YI-YSL

the animal pole view (Fig 1, 6) The deep cells derived

from the centre cells of 32-cell stage are covered by

other blastomeres completely, so they are also called

as the buried cells (Fig 2, 3) And the cells in the

top-most layer are called the enveloping layer (EVL) of the

blastoderm (Fig 2, 3)

Morula period (2 h 30 min^3 h 30 min)

From this period, the cleavage planes are not

pre-cisely traced any longer (Fig.1,7a, b) The blastomeres

become smaller and the cells of EVL are £attened (Fig

2, 4a) On the periphery, between the blastoderm and

the yolk, some marginal cells begin to fuse with the

yolk in close vicinity, yolk syncytial layer (YSL)

for-mation (Fig 2, 4b) As an organ unique to teleosts,

the YSL undergoes nuclear cleavages

unaccompa-nied of cytoplasmic cleavages, so the YSL becomes

multinucleate (Fig 2, 4b)

Blastula period (3 h 30 min^5 h 30 min)

High stage (3 h 30 min)

At this stage, successive cleavages of the top

blasto-meres result in the protuberating of blastoderm on the

animal pole, and it is called high blastula (Fig 1, 8a, b)

TheYSL extends underneath the blastoderm to form the

internal syncytium (I-YSL) (Fig 2, 5) The I-YSL makes

the blastoderm and yolk apart and it exists throughout

the whole embryonic development Between the

blasto-derm and I-YSL, some elliptical mini-cavities (Fig 2, 5)that will develop into the blastocoel, ¢rstly appear.Low (or sphere) stage (4 h 30 min)

As the animal^vegetal axis of the blastula ously shortens, the blastula becomes spherical withthe marginal blastoderm being smooth (Fig 1, 9a, b).The YSL begins to spread along the yolk mass andforms the external syncytium (E-YSL) (Fig 2, 6a) Alathy and elliptical blastocoel (Fig 2, 6b) forms be-tween the blastoderm and the I-YSL

continu-Gastrula period (5 h 30 min^12 h 30 min)Preliminary epiboly (5 h 30 min)

Led by E-YSL, the blastoderm begins to extend alongthe yolk mass towards the vegetal pole Moving alongthe curvature of the yolk sphere, the blastoderm be-comes thinner and it changes from a high protuber-ance to a multilayer cells uniform in thickness andshape (Fig 1, 10a)

25%-epiboly stage (6 h 30 min)The cell involution marks the onset of gastrula-tion This movement produces a thickening in theblastoderm marginal region, called germ ring(Fig 1, 11a) The germ ring (Fig 1, 11b) is formed bythe epiblast (upper) and the hypoblast (lower)(Fig 2, 7) The epiblast will give rise to the ectoderm,and the hypoblast will form the mesoderm and

np M

4a

nt My

My No

4b

dt

nr

No So

Figure 3 Notochord formation and neurulation in rock bream embryo [haematoxylin and eosin (H&E)] (1) The neuralplate ¢rstly appears: (a) bar 5 20mm; (b) bar 5 25 mm (2) The neural keel and notochord rudiment: (a) bar 5 50 mm; (b)bar 5 20mm (3) The neural rod: (a) bar 5 50 mm; (b) bar 5 25 mm (4) The neural tube and ‘stack of pennies’: (a)bar 5 20mm; (b) bar 5 25 mm dt, digestive tract; fp, £oor plate; Hyp, hypochord; I-YSL, internal syncytium; M, mesoderm;

mg, midgut; My, myotome; nk, neural keel; No, notochord; nr, neural rod; nt, neural tube; Pe, periderm; So, somite

the endoderm The Brachet’s cleft (Fig 2, 7, 8b), a

¢s-sure between the epiblast and hypoblast becomes

visible

50%-epiboly stage (7 h 30 min)

As a result of convergence, the germ ring becomes

somewhat broader and thicker at the point where

the embryonic axis will develop This triangular

re-gion is called the embryonic shield (Fig 1, 12a, b),

and it will spread towards the animal pole (Fig 2,8a) This stage is also a signal of the formation ofthe dorsal^ventral axis The upper of the embryonicshield will develop into the dorsal side of theembryonic body and the lower will develop intothe ventral side With the cells movement, thehypoblast involutes into the blastocoel, resulting

in the formation of archenteron and three germlayers (ectoderm, mesoderm and endoderm)(Fig 2, 8b)

No, notochord; op, optic vesicle; otp, otic placode; ot, optic tectum; ov, otic vesicle; r, rhombomere;Te, telencephalon;Y, yolk

Sof Y

haematox-75%-epiboly stage (9 h 30 min)

Epiboly continues to extend towards the vegetal pole,

and the blastoderm covers approximately 3/4 of the

yolk mass (Fig.1, 13) At this stage the E-YSL

degener-ates, while the I-YSL is still visible (Fig 3, 1a) The

neural plate derived from the ectoderm (Fig 1, 13, 3,

1) ¢rstly appears The EVL, on the surface of the

blas-toderm, continuously £attens and di¡erentiates into

the periderm (Fig 3, 1a) It consists of a single layer of

cells protecting the embryonic body

90%-epiboly stage (10 h 30 min)

At this stage, the blastoderm nearly covers the entire

yolk sphere In the vegetal pole, uncovered yolk mass

is called yolk plug (Fig 1, 14a) The neural keel

devel-ops by the folding of the neural plate (Fig 1, 14a, 3, 2)

The anterior of the neural keel begins to enlarge and

forms the brain primordium (Fig 4, 1, 2) This process

is considered cephalization In the area of brain

an-lage, the optic vesicle (Fig.1,14c, 4, 2,5,1) that consists

of ¢ve to six layers of squamous cells, the olfactory

vesicle (Fig 5, 1) that involves two layers of squamous

cells in the anterior of optic vesicle and the otic

pla-codes (Fig 4, 4) in the posterior of the optic vesicle allbecome visible in pairs At this stage, the notochordrudiment including a layer of round and big cells ap-pears ¢rstly (Fig 3, 2a) The earliest somitic furrow(Fig 6, 1) derived from the mesoderm develops in themiddle of the embryonic body

Complete epiboly (11h 30 min)

At this stage, movement of the epiboly is ¢nished andthe blastopore is closed (Fig.1,15) The neural keel de-velops into the neural rod (Fig.1,14b, 3, 3) The cells ofnotochord develop into two layers of spindle cellsfrom monolayer of round ones (Fig 3, 3a) The cellslying in close to the dorsal notochord begin to £atten

to form the £oor plate anlage (Fig 3, 3a) The brain ispartitioned into three segments: forebrain, midbrainand hindbrain (Fig 4, 3) The lumen of the epiphsis is

¢rst evident and the squmaous cells of the optic cle continue to increase At the end of this period, atail bud (Fig 1, 16a) has formed in the caudal end ofthe embryonic body It consists of a mass of undi¡er-entiated cells, which will give rise to the posteriortrunk and the tail In addition, the somites increase

vesi-to four pairs (Fig 6, 2)

ap-Segmentation period (12 h 30 min^20 h

30 min)

Notochord and neurulation

At13 h 30 min post fertilization, the cavitation of

neur-al rod results in the formation of the neurneur-al tube (Fig.3,

4), which is known as the ‘secondary neurulation’ The

hypochord, which consists of a monolayer of squamous

cells like as the £oor plate, earliest appears ventral to

the notochord (Fig 3, 4a) As the cells of notochord

va-cuolate and swell, the notochord shows itself as a‘stack

of pennies’ (Fig 3, 4a) in shape at 20 h 30 min

Brain

At 14 h 30 min, the forebrain is di¡erentiated into

anterior telencephalon and posterior diencephalon

The midbrain develops into the mesencephalon (Fig

4, 6) The cells in each segment of the brain are not

distinctly di¡erent At 17 h 30 min, cerebellum

an-lage (Fig 4, 7), a protuberance that is stained deeper

than other parts of brain, appears in the hindbrain

between the mesencephalon and rhombomeres The

brain is segmented into telencephalon, epiphysis,

diencephalon, mesencephalon, cerebellum andrhombomere until 19 h after fertilization (Fig 4,7).Eyes and ears

The two optic vesicles ¢rst appear at 90%-epibolystage (Fig.5,1) Then, the optic lumens continue to en-large (Fig 5, 2) At 15 h 30 min post fertilization, thelens placodes (Fig 5, 3) that involve a stu¡ of roundcells of tight arrangement ¢rst appear At 17 h 30 minafter fertilization, the cells of lens placodes increase to

¢ve layers At 20 h 30 min after fertilization, choroid

¢ssures (Fig.5, 4) are present in the ventral to the eyes

At 14 h 30 min post fertilization, the two otic codes are beginning to hollow (Fig 4, 8) By the hol-lowing of the otic placodes, the otic vesicles areformed with the surrounding of columnar cells (Fig

pla-4, 9) until 15 h 30 min after fertilization

Somites

At 12 h 30 min post fertilization, the somites increase

to six pairs (Fig 1, 16b) At 14 h 30 min after tion, the cells of somites elongate into the muscle ¢-bres (Fig 6, 3) At 15 h 30 min after fertilization thesomites develop into the myotomes, and they takes

mb spc

dt

Figure 8 Heart development in the rock bream embryo [haematoxylin and eosin (H&E)] (1) Heart primordium appears inthe ventral side of the midbrain, bar 5 20mm (2) Heart is divided into two chambers, bar 5 20 mm (3) Heart is di¡erentiatedinto four subdivisions: sinous venosus, atrium, ventricle and bulbous arteriosus, bar 5 50mm (4) Heart at the hatchingperiod, bar 5 20mm Atp, atrium primordium; Bap, bulbous arteriosus primordium; bc, blood cells; Ce, cerebellum; dt, diges-tive tract; fp, £oor plate; h, heart; hp, heart primordium; I-YSL, internal syncytium; mb, midbrain; Me, mesencephalon; No,notochord; r, rhombomere; sc, squamous cells; spc, spindle cells; svp, sinous venosus primordium;Vep, ventricle primordium

on a chevron shape (Fig 6, 4) The V points of

myo-tomes face the head of the embryonic body

Kup¡er’s vesicle (KV)

At 12 h 30 min post fertilization, the KV presents

it-self in the vicinity of tail bud (Fig 1, 16a) The KV, a

transient vacuole conserved in the teleosts embryos

consists of ventrally I-YSL and dorsally ciliated

co-lumnar cells (Fig 7, 1) At 17 h 30 min after

fertiliza-tion, it begins to degenerate gradually with the

in¢ltration of uncertain eosinophilic granule into

the vesicle (Fig.7, 2, 3, 4) Until 20 h 30 min after

ferti-lization, the KV vanishes completely (Fig 1, 17)

Heart

At 13 h 30 min post fertilization, the heart

primor-dium (Fig 8, 1), an ellipsoidal tube derived from the

mesoderm, appears in the ventral side of the

mid-brain It consists of I-YSL exteriorly and a stu¡ of

spindle cells interiorly At 17 h 30 min after

fertiliza-tion, the heart is divided into two chambers, and the

spindle cells turn into the squamous cells (Fig 8, 2)

At 19 h after fertilization, the heart is segmented into

four subdivisions: the primordium of sinous venosus,

atrium, ventricle and bulbous arteriosus (Fig 8, 3) At

20 h 30 min post fertilization, the heart is beginning

to beat and the blood (Fig 8, 4) starts to circulatethrough a closed set of channels; however, the peri-cardial cavity is not yet prominently in£ated

Gut and kidneyAt13 h 30 min after fertilization, in the ventral side ofthe embryonic body, the digestive tract primordium(Fig 9, 1) derived from the endoderm, is ¢rst visiblewith two layers of squamous cells At 19 h post fertili-zation, the lumen of digestive tract (Fig 9, 2) with thesquamous epithelium becomes ¢rst visible in theanterior gut It is called as foregut At 20 h 30 minafter fertilization, the midgut with the monolayer cu-bic epithelium appears in the middle of digestive tract(Fig 9, 3)

At14 h 30 min post fertilization, the pronephric bule primordium (Fig 9, 2), which involves two layers

tu-of squamous cells, ¢rst appears between the gut andthe notochord At 15 h 30 min after fertilization, thelumen (Fig 9, 3) of the pronephric tubule with mono-layer squamous epithelium is earliest visible

At the onset of pigmentation, melanophores are

¢rst evident in the dorsolateral side of the embryo(Fig 4, 5) at 17 h 30 min post fertilization The ¢rstprotrusion of the unpaired embryonic ¢n-fold (Fig 7,

2

No

dtptp

cc

Mel

Figure 9 Gut and kidney development in the rock bream embryo (bar 5 20mm) (1) The digestive tract primordium, matoxylin and eosin (H&E) (2) The pronephric tubule primordium, Mallory’s trichrome stain (3) The lumen of pronephrictubule and digestive tract, H&E (4) The midgut and hindgut, H&E cc, columnar cells; cu, cubic epithelium; dt, digestivetract; dtp, digestive tract primordium; hg, hindgut; Mel, melanophores; mg, midgut; My, myotome; No, notochord; pt, pro-nephric tubule; ptp, pronephric tubule primordium; sc, squamous cells

hae-2) develops in the tail bud at 19 h after fertilization

and then extends along the trunk (Fig.7, 3)

Hatching period (22^25 h)

As the choroid ¢ssures widen, the corneas of eyes,

which consist of monolayer cubic cells, develop in

the site (Fig 5, 5) The unpaired embryonic ¢n-fold

ex-pands to surround the entire body with the exception

of the head The lumen in the posterior of digestive

tract becomes evident with monolayer columnar

cells, which marks the formation of the hindgut (Fig

9, 4) As the embryonic body coils 4/5 of the yolk sac

at 25 h after fertilization, embryo begins to hatch out

(Fig 1, 18, 19)

Discussion

In teleosts, the fertilized eggs show a discoidal

mero-blastic cleavage pattern However, the orientation of

the cleavage furrow is not identical among teleosts

In rock bream, the ¢rst cleavage furrow is vertically

oriented, as is usual until the ¢fth cleavage between

the 16- and 32-cell stages At this stage, the

horizon-tal cleavage ¢rst occurs and the uneven cleavage

makes the blastoderm partially strati¢ed with single

layer on the periphery and two cell layers in the

cen-tre It is just the same as medaka (Iwamatsu 2004)

occurs in the third cleavage between four- and

eight-cell stages In zebra¢sh (Kimmel et al 1995),

ti-lapia Oreochromis niloticus (Morrison et al 2001),

Atlantic cod Gadus morhua (Hall et al 2004) and

cy-prinoid Anabarilius grahami (Ma, Pan,Wei, Li, Li,Yang

& Mao 2008), it occurs initially in the sixth cleavage

between 32- and 64-cell stage In summer £ounder

(Martinez & Bolker 2003), it appears much later,

be-tween 64- and 128-cell stage in the seventh cleavage

At high stage of rock bream embryo, the blastocoel

develops between the blastoderm and the I-YSL

Blas-tocoel was also observed in the annual ¢sh A viarius,

and was called as the segmentation cavity (Arezo

et al 2005) However, the blastocoel does not appear

in all teleosts In zebra¢sh (Kimmel et al 1995) and

cyprinoid A grahami (Ma et al 2008), the blastulas

were solid without blastocoele formation

The somitegenesis in teleost embryogenesis

usual-ly shows species speci¢city At 90%-epibousual-ly stage in

rock bream, the earliest somitic furrow that derived

from the mesoderm forms in the middle of embryo

This is the same as the summer £ounder (Martinez

& Bolker 2003), in which the somite was earliest

dis-cernible at 75^95% epiboly stage Similarly, in mon Japanese conger, six to eight indistinct somiteswere ¢rst visible at 80%-epiboly stage (Horie et al.2002) In Atlantic cod (Hall et al 2004), the ¢rst so-mite appears at 45%-epiboly stage Nevertheless, inzebra¢sh, sunbleak, medaka, striped snakehead andcyprinoid, the ¢rst somite cannot be viewed untilthe completion of epiboly (Kimmel et al 1995; Iwa-matsu 2004; Pinder & Gozlan 2004; Marimuthu &Hani¡a 2007; Ma et al 2008)

com-The KV is a transitory organ peculiar to early ost embryos In rock bream, with the completion ofepiboly, the KV presents itself in the vicinity of tailbud It is recognized as a transient vacuole conserved

tele-in the teleost embryo (Kimmel et al 1995; Morrison

et al 2001; Horie et al 2002; Iwamatsu 2004; Hall

et al 2004; Arezo et al 2005) From 17 h 30 min to

22 h after fertilization, the KV had gradually erated with penetration of eosinophilic granules un-til it disappeared completely, but the granules, whichmay origin from I-YSL or yolk mass, were not identi-

degen-¢ed The process of degeneration of the KV had notbeen reported in other species The KV was ¢rst de-scribed by Kup¡er in 1868 However, its function hasnot been understood completely In zebra¢sh, someauthors reported that the KV initiated left^right de-velopment by establishing a directional £uid £owand the £ow rate was calculated (Essner, Vogan,Wagner, Tabin, Yost & Brueckner 2002; Essner,Amack, Nyholm, Harris & Yost 2005; Kreiling, Prab-hat,Williams & Creton 2007) In medaka, Hojo,Taka-shima and Kobayashi (2007) reported the genecharon was the earliest asymmetric gene, and it could

be used for a gene marker to analyse the early pattern

of the £uid £ow

Like as other teleosts (Kimmel et al 1995; Hall et al.2004; Pinder & Gozlan 2004; Marimuthu & Hani¡a2007) at hatching period, the heart of rock bream isdi¡erentiated into four segments And at 20 h 30 minafter fertilization, the heart in rock bream embryo be-gins to beat at 23.5 0.5 1C This process in sunbleak(Pinder & Gozlan 2004) took 27 h 15 min with the in-cubation temperature of 24.7 1.9 1C In zebra¢sh(Kimmel et al 1995), it needed 24 h at 28.5 1C, and instriped snakehead (Marimuthu & Hani¡a 2007), only

20 h at 29 1 1C However, other species with lowerincubation temperature may need more time, such asAtlantic cod (Hall et al 2004) It took 200 h when theheart begins to beat at 7 0.2 1C Nevertheless, thehigher incubation temperature does not always makethe heartbeat earlier In medaka (Iwamatsu 2004),the heart began to beat even 50 h after fertilization

at 26 1 1C Above results indicated that the

incuba-tion temperature is an important factor, but not the

exclusive one for embryogenesis

At hatching stage, most teleosts show an

undi¡er-entiated digestive tract without any aperture to the

exterior and the yolk sac supplies the nutrition to

the embryo In rock bream, the digestive tract is a

straight tubule with incomplete lumen Similar

re-sult was reported in Atlantic cod (Hall et al 2004),

but its deep folds in the gut are visible However, the

digestive tract of zebra¢sh (Kimmel et al.1995) is more

developed at hatching Its mouth is observable and

open though without gaping Even in tilapia, the

oe-sophagus and the stomach have di¡erentiated

al-ready (Morrison et al 2001) The early kidney, which

is called pronephros, forms between the gut and the

notochord with simple squamous epithelium in rock

bream However, in the Atlantic cod, even the

glomer-uli appeared in pronephros at the hatching stage

(Hall et al 2004)

Histological and morphological studies are

indis-pensable to illuminate the developmental

mechan-ism of teleosts In the present study, we observed the

histo-morphological characteristic of rock bream

embryogenesis, and it would be useful for further

un-derstanding the early ontogeny However, for

investi-gating whole early development, further studies on

rock bream larvae and juveniles to gain more

infor-mation of developmental biology in the future are

needed

Acknowledgments

This work was supported by the Knowledge

Innova-tion Program of Chinese Academy of Sciences

(KSCX2-YW-N-47-08), Funds of Promoting

Transfor-mation of National Agriculture Achievements

(05EFN2166000453) and the Science-Technology

R&D Project of Qingdao, China (05-1-HY-79).We wish

to express our sincere thanks to Z H Zhang for her

assistance in raising the adults of rock bream and G

Z Wu for his expert advice in preparing the

manu-script

References

Arezo M.J., Pereiro L & Berois N (2005) Early development

in the annual ¢sh Cynolebias viarius Journal of Fish

Biol-ogy 66, 1357^1370.

Chang K.M., Mao J.P., Wu J.F & Zhang K.J (2005) Study on

the embryo and larvae ¢sh development of Oplegnathus

fasciatus Journal of Shanghai Fisheries University 14, 401^405.

ChoY.S., Choi B.N., Kim K.H., Kim S.K., Kim D.S., Bang I.C & NamY.K (2006) Di¡erential expression of Cu Zn superox- ide dismutase mRNA during exposures to heavy metals

in rockbream Oplegnathus fasciatus Aquaculture 253, 667^679.

Choi S.K., Kwon S.R., Nam Y.K., Kim S.K & Kim K.H (2006) Organ distribution of red sea bream iridovirus (RSIV) DNA in asymptomatic yearling and ¢ngerling rock bream (Oplegnathus fasciatus) and e¡ects of water temperature

on transition of RSIV into acute phase Aquaculture 256, 23^26.

Essner J.J., Amack J.D., Nyholm M.K., Harris E.B & Yost H.J (2005) Kup¡er’s vesicle is a ciliated organ of asymmetry

in the zebra¢sh embryo that initiates left-right development of the brain, heart and gut Development

132, 1247^1260.

Essner J.J., Vogan K.J., Wagner M.K., Tabin C.J., Yost H.J & Brueckner M (2002) Conserved function for embryonic nodal cilia Nature 418, 37^38.

Franz L & Pierpaolo P (2004) Egg quality determination in the gilthead seabream, Sparus aurata, with biochemical parameters Aquaculture 237, 443^459.

Furlani D.M & Ruwald F.P (1999) Egg and larval ment of laboratory-reared striped trumpeter Latris linea- tea (Forster in Bloch and Schneider 1801) (Percoidei: Latridiidae) from Tasmanian waters New Zealand Journal

develop-of Marine and Freshwater Research 33, 153^162.

Hall T.E., Smith P & Johnston I.A (2004) Stages of nic development in the the Atlantic Cod Gadus morhua Journal of Morphology 259, 255^270.

embryo-Hojo M.,Takashima S & Kobayashi D (2007) Right-elevated expression of charon is regulated by £uid £ow in medaka Kup¡er’s vesicle Development Growth & Di¡erentiation 49, 395^405.

Horie N., Utoh T.,YamadaY., Okamura A., Zhang H., Mikawa N., Akazawa A.,Tanaka S & Poka H (2002) Development

of embryos and larvae in the common Japanese conger Conger myriaster Fisheries Science 68, 972^983 Iwamatsu T (2004) Stages of normal development in the medaka Oryzias latipes Mechanisms of Development 121, 605^618.

Jung S.J & Oh M.J (2000) Iridovirus-like infection ciated with high mortalities of striped beakperch, Opleg- nathus fasciatus (Temminck et Schlegel), in southern coastal areas of the Korean peninsula Journal of Fish Dis- eases 23, 223^226.

asso-Kim Y.I., Ha Y.M., Ahn S.J., Nam Y.K., asso-Kim K.H & asso-Kim S.K (2007) Production and characterization of polyclonal antibody against recombinant ORF 049L of rock bream (Oplegnathus fasciatus) iridovirus Process Biochemistry

42, 134^140.

Kimmel C.B., Ballard W.W., Kimmel S.R., Ullmann B & ling T.F (1995) Stages of embryonic development of the zebra¢sh Developmental Dynamics 203, 253^310.

Schil-Kreiling J.A., Prabhat,Williams G & Creton R (2007)

Analy-sis of Kup¡er’s vesicle in zebra¢sh embryos using a cave

automated virtual environment Developmental

Dy-namics 236, 1963–1969.

Lee S.J., Kim S.M., Kim S.K., Nam Y.K., Kim D.S & Kim K.H.

(2004) Modulation of zymosan- and phrbol

ester-stimu-lated respiratory burst of rock bream (Oplegnathus

fascia-tus) phagocytes by staurosporine, genistein and sodium

orthovanadate Aquaculture 237, 1^7.

Lim S.J & Lee K.J (2009) Partial replacement of ¢sh meal by

cottonseed meal and soybean meal with iron and phytase

supplementation for parrot ¢sh Oplegnathus fasciatus.

Aquaculture 290, 283^289.

Liu X.Z., XuY.J.,WangY.Y., LuY.Q & Qu J.Z (2008) Characters

of development and growth of early life stages of the rock

bream Oplegnathus fasciatus Acta Zoologica Sinica 54,

332^341.

Ma L., Pan X.F., Wei Y.H., Li Z.Y., Li C.C.,Yang J.X & Mao B.Y.

(2008) Embryonic stages and eye-speci¢c gene expression

of the local cyprinoid ¢sh Anabarilius grahami in Fuxian

Lake, China Journal of Fish Biology 73, 1946^1959.

Marimuthu K & Hani¡a M.A (2007) Embryonic and larval

development of the striped snakehead Channa striatus.

Taiwania 52, 84^92.

Martinez G.M & Bolker J.A (2003) Embryonic and larval

staging of summer £ounder (Paralichthys dentatus)

Jour-nal of Morphology 255, 162^176.

Morrison C.M., Miyake T & Wright J.R (2001) Histological

study of the development of the embryo and early larva of

Oreochromis niloticus Journal of Morphology 247,172^195.

Nam Y.K., Cho Y.S., Choi B.N., Kim K.H., Kim S.K & Kim D.S (2005) Alteration of antioxidant enzymes at the mRNA level during short-term starvation of rock- bream Oplegnathus fasciatus Fisheries Science 71, 1385^1387.

Pinder A.C & Gozlan R.E (2004) Early ontogeny of

sunble-ak Journal of Fish Biology 64,762^775.

Quan H.F & Xiao Z.Z (2007) Research on the technique for arti¢cial breeding of Oplegnathus fasciatus Journal of Oceanography inTaiwan Strait 26, 295^300.

Rodda K.R & Seymour R.S (2008) Functional morphology

of embryonic development in the Port Jackson shark erodontus portusjacksoni (Meyer) Journal of Fish Biology

Het-72, 961^984.

Shan X.J., Quan H.F & Dou S.Z (2008) E¡ects of delayed

¢rst feeding on growth and survival of rock bream Oplegnathus fasciatus larvae Aquaculture 277, 14^23.

Wang X.J., Kim K.W., Bai S.C., Huh M.D & Cho B.Y (2003) E¡ects of the di¡erent levels of dietary vitamin C

on growth and tissue ascorbic acid changes in parrot ¢sh (Oplegnathus fasciatus) Aquaculture 215, 203^211.

Xiao Z.Z., Zheng J.,Yu D.D & Li J (2008) Developmental acters at the early stages of the Japanese Parrot¢sh (Opleg- nathus fasciatus) Marine Sciences 32, 25^30.

char-Yoshikoshi K & Inoue K (1990) Viral nervous necrosis in hatchery-reared larvae and juveniles of Japanese parrot-

¢sh, Oplegnathus fasciatus (Temminck et Schlegel) Journal

of Fish Diseases 13, 69^77.

Cloning, expression and sequence analysis of

Indian isolate

Mudagandur S Shekhar1, Pramoda K Sahoo2, Manickam Dillikumar1& Abhilipsa Das2

1 Genetics and Biotechnology Unit, Central Institute of Brackishwater Aquaculture, Chennai, India

2 Central Institute of Freshwater Aquaculture, Fish Health Management Division, Bhubaneswar, India

Correspondence: Dr M S Shekhar, Genetics and Biotechnology Unit, Central Institute of Brackishwater Aquaculture,75 Santhome High Road, R A Puram, Chennai, India E-mail: msshekhar@hotmail.com

Abstract

RNA-dependent RNA polymerase (RdRp), B2 and

capsid genes of Macrobrachium rosenbergii

noda-virus (MrNV) of Indian isolate were polymerase

chain reaction ampli¢ed, cloned and sequenced

Expression of the MrNV fusion recombinant

pro-teins of RdRp (44.5 kDa), B2 (32.2 kDa) and capsid

(58.4 kDa) was con¢rmed by Western blot analysis

using anti-His mouse monoclonal antibodies

Poly-clonal antibodies speci¢c to puri¢ed recombinant

MrNV capsid protein showed speci¢city against the

capsid protein byWestern blot The protein sequence

analysis of the partial RdRp gene of MrNV revealed

the signature sequence along with the conserved core

residues of the catalytic domain and indicated the

pre-sence of active sites, metal ion-binding site and nucleic

acid-binding site residues The Indian isolate of MrNV

showed high RdRp and capsid gene sequence

homol-ogy with the other MrNV geographical isolates

How-ever, the Belize isolate was found to be the most

distinct among the di¡erent geographical prawn

no-davirus isolates due to the host speci¢city Secondary

structure prediction analysis of the MrNV capsid

pre-dicted it to be a DNA-binding protein consisting ofa

helix (22.91%), extended strand (24.80%), b turn

(5.39%) and random coil (46.90%) regions

Keywords: Macrobrachium rosenbergii nodavirus,

RNA-dependent RNA polymerase (RdRp), B2,

cap-sid, recombinant proteins

Introduction

Incidence of white tail disease (WTD) in

Macrobra-chium rosenbergii has been reported from French

West Indies (Arcier, Herman, Lightner, Redman,Mari & Bonami 1999), China (Qian, Shi, Zhang, Cao,Liu, Li, Xie, Cambournac & Bonami 2003), India (Sa-hul Hameed, Yoganandhan, Sri Widada & Bonami2004; Shekhar, Azad & Jithendran 2006), Taiwan(Hsieh,Wu,Tung,Tu, Lo, Chang, Chen, Hsieh & Tsai2006), Thailand (Yoganandhan, Leartvibhas, Sri-wongpuk & Limsuwan 2006) and very recentlyfrom Australia (Owens, La Fauce, Juntunen, Haya-kijkosol & Zeng 2009) The causative agents of thedisease have been identi¢ed as large virus M rosen-bergii nodavirus (MrNV) and extra small virus(XSV) Macrobrachium rosenbergii nodavirus is anicosahedral, non-enveloped virus of 26^27 nm size

It contains two pieces of ssRNA, of 3.0 kb (RNA-1)and 1.2 kb (RNA-2) (Qian et al 2003) RNA-1 ofMrNV is suggested to encode for RNA-dependentRNA polymerase (RdRp), whereas the capsid, which

is composed of a single polypeptide of 43 kDa, is coded by RNA-2 of MrNV (Bonami, Shi, Qian & SriWidada 2005) In addition, a subgenomic RNA tran-scribed from the 30end of RNA1, termed RNA3, en-coding a protein called B2 has been reported inbetanodaviruses (Sommerset & Nerland 2004),which is suggested to play a role as an RNA interfer-ence antagonist (Fenner, Thiagarajan, Chua &Kwang 2006) The sequence of B2 gene has been re-ported for MrNV (Sri Widada, Durand, Cambournac,Qian, Shi, Dejonghe, Richard, & Bonami 2003) indi-cating the presence of putative B2 gene in MrNV.The smaller virus, XSV is an icosahedral, non-en-veloped virus of 14^16 nm size and possesses a linearssRNA genome of 0.9 kb Because XSV does not pos-sess gene encoding for RdRp (Sri Widada & Bonami2004), it is suggested to be a satellite virus, which is

en-dependent on MrNV for its replication (Qian et al.

2003; Bonami et al 2005) However, the exact

rela-tionship between the two viruses is still not clear

(Zhang,Wang ,Yuan, Li, Zhang, Bonami & Shi 2006)

In the present study cloning, expression of viral

re-combinant proteins and sequence analysis have been

carried out to characterize the MrNV of Indian

iso-late The sequence analysis between di¡erent viral

geographical isolates, indicated host-speci¢c

se-quence variation in the viral capsid

Materials and methods

Ampli¢cation and cloning ofMrNV RdRp, B2

and capsid genes

The polymerase chain reaction (PCR) product of

MrNV RdRp gene (859 bp) was obtained in our

ear-lier study (Shekhar et al 2006) using the upstream

primer 50CCACGTTCTTAGTGGATCCT30 and the

downstream primer 50CGTCCGCCTGGTAGTTCC30

speci¢c to RNA-1 as reported by Sri Widada et al

(2003) The PCR product was sequenced and the

se-quence has now been assigned GenBank accession

number DQ146969 This PCR product in the present

study was reampli¢ed with a new set of internal

pri-mers containing restriction sites for cloning The

up-stream primer 50CGGCCATGGAAGTCCGCCGA30

containing NcoI restriction site and the downstream

primer 50GCCAAGCTTTTACCACGTTCTTAG 30

con-taining HindIII restriction site were designed to

ob-tain a 738 bp PCR product (The restriction sites are

underlined in the primer sequences) The PCR cycle

consisted of initial denaturation at 94 1C for 2 min

followed by 30 cycles of denaturation at 92 1C for

1min, annealing at 55 1C for 1min and extension at

72 1C for 1min with a ¢nal cycle of 10-min extention

at 72 1C The PCR product was gel extracted using a

QIAquick gel extraction kit (Qiagen, Hilden,

Ger-many) and restriction digestion was carried out with

NcoI and HindIII restriction enzymes The PCR

pro-duct was then ligated to pET32a (1) vector

(Nova-gen, Darmstadt, Germany) linearized with NcoI and

HindIII restriction enzymes The ligated mixture was

transformed into competent DH5a cells Plasmids

were isolated from the transformed colonies followed

by restriction enzyme analysis with NcoI and HindIII

restriction enzymes to con¢rm the clones for the

pre-sence of the insert

Polymerase chain reaction ampli¢cation for

MrNV B2 gene was carried out by designing the

pri-mers based on the GenBank accession number

NC_005094 Ampli¢cation to obtain 402 bp PCR duct was carried out using the upstream primer

pro-50ATGCAGTGGACGAACGTCAAT30 and the stream primer 50TTACCACGTTATGAGGTCGC30.Full-length PCR ampli¢cation for MrNV capsidgene was carried out by designing the primers speci-

down-¢c to RNA-2, based on the GenBank accession ber NC_005095 Ampli¢cation to obtain 1116 bp PCRproduct was carried out using the upstream primer

num-50ATGGCTAGAGGTAAACAAAATTCTAA30and thedownstream primer 50CTAGTTATTGCCGACGATAGCTCTGA30 The primers for MrNV B2 and capsidgenes were modi¢ed to contain NcoI restriction site

in the upstream primer and HindIII restriction site

in the downstream primer for cloning The PCR cycleconditions and the cloning procedures for MrNV B2and capsid genes were same as followed for RdRpgene of MrNV

Dot blot hybridizationThe 859 bp PCR product of MrNV RdRp gene was usedfor generating probe by digoxigenin (DIG) randomprimed DNA labelling using a DIG DNA labelling anddetection kit (Roche, Mannheim, Germany) followingthe manufacturer’s protocol One microlitre of recom-binant plasmids-containing RdRp gene was isolatedfrom transformed cells and spotted on nylon mem-brane for dot blot analysis One microlitre of 859 PCRproduct of MrNV and pET32a (1) plasmid DNA werealso spotted on the same nylon membrane as a posi-tive and negative control respectively The DNA spots

on the membrane were ¢xed by cross linking with UVlight Prehybridization, hybridization and detectionwith NBT/BCIP for the dot blot experiments were per-formed as per the protocol described in the kit

Expression and puri¢cation of recombinantMrNV RdRp, B2 and capsid proteins

Plasmids isolated from the positive clones of DH5acells containing the gene inserts for MrNV RdRp, B2and capsid were transformed into competentBL21(DE3) pLysS cells The positive clones wereinitially screened by colony PCR followed by NcoIand HindIII restriction enzyme analysis to con¢rmthe presence of the viral gene inserts in the vectorDNA The positive colonies were grown overnight inLuria Bertani (LB) broth containing 100mg mL 1ampicillin The overnight culture was inoculated(1:10) a dilution of fresh LB broth containing

100mg mL 1ampicillin and incubated with shaking

till the optical density (OD600) reached 0.4 to obtain

the cells at mid log phase Isopropylb-D

-thiogalacto-pyranoside (IPTG) was added at 1mM ¢nal

concen-tration and the cells were harvested 4 h post IPTG

in-duction The harvested cells were lysed with a

bacterial cell lysis reagent (Bangalore Genei,

Banga-lore, India) by incubating the cell suspension at room

temperature with shaking for 20 min followed by

so-nication on ice with three 5 s pulses Expression of

recombinant protein was analysed by SDS-PAGE,

which included uninduced bacterial culture as a

ne-gative control

pET32a (1) vector has six histidine residues,

which enable to express and purify recombinant

fu-sion protein carrying histidine tag (His.Tag)

Puri¢ca-tion of the MrNV recombinant proteins was carried

out using ProBond puri¢cation system (Invitrogen,

Carlsbad, CA, USA) following the manufacturer’s

in-structions Brie£y, IPTG-induced bacterial cells were

harvested from 50 mL of culture The cell pellet was

resuspended in 8 mL of guanidium lysis bu¡er (6 M

guanidine HCL, 20 mM sodium phosphate, pH 7.8,

500 mM NaCl) with gentle rocking at room

tempera-ture for 10 min The bacterial cell suspension was

so-nicated on ice with three 5 s pulses followed by

centrifugation at 11000 g for 15 min The

recombi-nant proteins were puri¢ed by passing the

superna-tant of the bacterial lysate through a ProBond resin

column

Polyclonal antibody production

The puri¢ed recombinant MrNV capsid protein

(58.4 kDa) protein was used for immunization in

New Zealand white rabbit for raising polyclonal

hy-per immune serum The rabbits were ¢rst immunized

with 300mg of recombinant protein emulsi¢ed with

Freund’s complete adjuvant, followed by two booster

doses of 200mg each of protein emulsi¢ed with

Freund’s incomplete adjuvant at 30 and 45 days

Blood was collected 20 days after the last booster

dose and the serum was separated The

immunoglo-bulin G fraction was puri¢ed by protein A a⁄nity

chromatography

Western blot

Western blot analyses of MrNV RdRp, B2 and capsid

recombinant proteins were carried out by the

stan-dard procedure using a semidry western blot

appara-tus (Hoe¡er, Holliston, MA, USA) The recombinantfusion proteins of MrNV (RdRp, B2 and capsid) ex-pressed from the pET32a (1) expression vector weredetected using the Western Breeze chromogenic wes-tern blot immunodetection kit, with alkaline phos-phatase-conjugated anti-His mouse monoclonalantibodies (1:10 000 dilution) (Invitrogen) The mem-branes were developed using BCIP/NBT as a sub-strate.Western blot was also carried out to check thespeci¢city of the polyclonal antibodies raised againstrecombinant MrNV capsid protein The membranestransferred with MrNV capsid protein were incu-bated with anti MrNV capsid primary antibodies(1:20 000 dilution) followed by horseradish peroxi-dase (HRP)-conjugated mouse anti-rabbit secondaryantibody (Bangalore Genei) at 1:20 000 dilution.The membranes were developed using TMB as asubstrate

Sequence analysisNucleotide and amino acid sequence analysis andphylogenetic trees were constructed using AdvancedGENEBEE CLUSTALW1.83 (http://www.genebee.msu.ru/clustal/advanced.html).PREDICTPROTEIN(http://www.predictprotein.org) andSOPMA, a web-based software,was used for the prediction of protein secondarystructure (http://npsa-pbil.ibcp.fr) Conserved do-main database search was carried out using RPS ^BLAST (http://www.ncbi.nlm.nih.gov/Structure/cdd/cddsrv.cgi)

Results and discussionPositive clones of the transformed DH5a cells con-taining recombinant plasmids of pET32a (1) withMrNV gene inserts, on PCR ampli¢cation revealedampli¢ed products in the expected size for RdRp(738 bp), B2 (402 bp) and capsid genes (1116 bp) asshown in Fig 1 These clones were further con¢rmed

by NcoI and HindIII restriction enzyme analysis Therecombinant plasmids released the respective geneinserts of MrNV as shown in Fig 2 Dot blot hybridi-zation using DIG DNA labelling and detection kit,could successfully detect positive clones containingrecombinant plasmids of pET32a (1) with MrNVRdRp gene insert (Fig 3)

In the present study, MrNV RdRp, B2 and capsidgenes have been expressed as recombinant fusionproteins in Escherichia coli Expression of MrNV re-combinant proteins of RdRp (44.5 kDa), capsid

(58.4 kDa) and B2 (32.2 kDa) genes was analysed by

SDS-PAGE, which included uninduced bacterial

cul-ture as negative controls Puri¢cation of recombinant

proteins carried out using a⁄nity column revealed

a single protein band in the expected size range

(Fig 4a^c).Western blot using anti-His mouse clonal antibodies con¢rmed the expression of viralrecombinant proteins, as shown in Fig 5 The poly-clonal antibodies raised against MrNV capsid re-vealed speci¢c immunoreactivity with the expressedand the puri¢ed recombinant protein (Fig 6)Viral RdRp plays an important role in RNA genomereplication, which is an essential step in the patho-genesis of many RNA viruses Not much information

mono-is available regarding the mode of infection and virusreplication in M rosenbergii infected with MrNV.Macrobrachium rosenbergii nodavirus infection in M.rosenbergii was shown to be located in striated muscletissues by in situ hybridization technique (Sri Widada

et al 2003) and in connective tissue by transmissionelectron microscopy (Qian et al 2003) To our knowl-edge, a sandwich enzyme-linked immunosorbent as-say (Romestand & Bonami 2003) and monoclonalantibodies-based triple antibody sandwich enzyme-linked immunosorbent assay (Qian, Liu, Jianxiang

&Yu 2006) are the only immuno-based techniques,which have been developed to diagnose WTD In boththe techniques, antibodies were raised against puri-

¢ed suspension of virus Expression of recombinantRdRp, B2 and capsid proteins of MrNV in E coli sys-tem is an easy and inexpensive technique for produ-cing speci¢c antigen, which can be subsequentlypuri¢ed and used for raising antiserum With theavailability of antiserum against MrNV proteins, itwill be possible to carry out histopathological investi-gations for in depth examination of MrNV infection

in di¡erent tissues of M rosenbergii The use of serum and DIG probe to study shrimp viral pathogen-

anti-Figure 1 PCR-ampli¢ed products of MrNV capsid, RdRp

and B2 genes Lane 1 l DNA EcoRI/HindIII molecular

weight marker Lane 2 MrNV capsid (1116 bp) Lane 3

MrNV RdRp (738 bp) Lane 4 MrNV B2 (402 bp) Lane 5

100 bp molecular weight marker PCR, polymerase chain

reaction; MrNV, Macrobrachium rosenbergii nodavirus;

RdRp, RNA-dependent RNA polymerase

Figure 2 NcoI and HindIII restriction digestion of

pET32a (1) recombinant plasmids containing MrNV

gene inserts (A) Lane 1.l DNA EcoRI/HindIII molecular

weight marker Lane 2 MrNV gene insert RdRp (738 bp)

Lane 3 PCR product MrNV RdRp (738 bp) (B) Lane 1.l

DNA EcoRI/HindIII molecular weight marker Lane 2

MrNVgene insert B2 (402 bp) Lane 3 PCR product MrNV

B2 (402 bp) (C) Lane 1.l DNA EcoRI/HindIII molecular

weight marker Lane 2 MrNV gene insert capsid

(1116 bp) Lane 3 PCR product MrNV capsid (1116 bp)

PCR, polymerase chain reaction; MrNV, Macrobrachium

rosenbergii nodavirus; RdRp, RNA-dependent RNA

poly-merase

Figure 3 Screening of recombinant plasmids containingMrNV RdRp gene by dot blot hybridization The PCRampli¢ed product (859 bp) of MrNV used as a positive con-trol and pET32a (1) plasmid DNA as a negative controlare indicated PCR, polymerase chain reaction; MrNV,Macrobrachium rosenbergii nodavirus; RdRp, RNA-depen-dent RNA polymerase

esis and tissue tropism has been reported For

exam-ple, a DIG-labelled white spot syndrome virus

(WSSV)-speci¢c probe was used to detect the virus

in haemocytes from WSSV-infected cray¢sh

(Jirava-nichpaisal, Sricharoen, S˛derhll & S˛derhll 2006)

Digoxigenin-labelled probe directed against RdRp ofvirus involved in monodon slow growth syndrome

in cultured black tiger shrimp (Penaeus monodon) hasbeen reported for use in pathological investigations

by in situ hybridization (Sritunyalucksana, takan , Boon-nat, Withyachumnarnkul & Flegel2006) In case of WTD, DIG-labelled DNA probesbased on MrNV RNA-2 as DNA template, have been

Apisawe-Figure 4 Expression of MrNV RdRp, capsid and B2 recombinant proteins in Escherichia coli strain BL21(DE3)pLysS (a)Lane 1 Protein molecular weight marker Lane 2 Uninduced cells (negative control) Lane 3 IPTG-induced cells showingprotein expression for MrNV RdRp Lane 4 Puri¢ed MrNV RdRp recombinant protein (44.55 kDa) (b) Lane 1 Proteinmolecular weight marker Lane 2 Uninduced cells (negative control) Lane 3 IPTG-induced cells showing protein expres-sion for MrNVcapsid Lane 4 Puri¢ed MrNVcapsid recombinant protein (58.4 kDa) (c) Lane 1 Protein molecular weightmarker Lane 2 Uninduced cells (negative control) Lane 3 IPTG-induced cells showing protein expression for MrNV B2.Lane 4 Puri¢ed MrNV B2 recombinant protein (32.2 kDa) MrNV, Macrobrachium rosenbergii nodavirus; RdRp, RNA-de-pendent RNA polymerase; IPTG, isopropylb-D- thiogalacto-pyranoside

Figure 5 Western blot analysis of MrNV RdRp, B2 and

capsid recombinant proteins using alkaline

phosphatase-conjugated anti-His mouse monoclonal antibodies Lane 1:

Coloured protein marker Lane 2: MrNV RdRp

IPTG-induced BL21(DE3) pLysS cells Lane 3: MrNV RdRp

uninduced BL21(DE3) pLysS cells Lane 4: MrNV B2

IPTG-induced BL21(DE3) pLysS cells Lane 5: MrNV B2

unin-duced BL21(DE3) pLysS cells Lane 6: MrNVcapsid

IPTG-in-duced BL21(DE3) pLysS cells Lane 7: MrNV capsid

uninduced BL21(DE3) pLysS MrNV, Macrobrachium

rosen-bergii nodavirus; RdRp, RNA-dependent RNA polymerase;

IPTG, isopropylb-D- thiogalacto-pyranoside

Figure 6 Western blot analysis of the polyclonal dies raised against recombinant MrNV capsid protein.Lane1 MrNVcapsid puri¢ed protein Lane 3 MrNVcapsidIPTG-induced BL21(DE3) pLysS cells Lane 3 MrNV cap-sid uninduced BL21(DE3) pLysS cells MrNV, Macrobra-chium rosenbergii nodavirus

antibo-successful in the detection of MrNV infection in

tis-sues of WTD-diseased prawns by in situ hybridization

(Sri Widada et al 2003; Hsieh et al 2006; Wang,

Chang,Wen , Shih & Chen 2008).Very limited reports

are available on tissue tropism associated with MrNV

pathogenesis Arcier et al (1999) by histopathological

examination reported apparent viral inclusions, but

much less conspicuous, in connective tissue of the

subcutis and rarely present in the gills in the infected

postlarvae of M rosenbergii prawn The use of in situ

hybridization technique for histopathological studies

by Sri Widada et al (2003) revealed no positive

reac-tions for the presence of MrNV viral RNA in gill and

hepatopancreatic tissues of M rosenbergii Basophilic

cytoplasmic inclusion bodies in the musculature and

hepatopancreas has been reported by Hsieh et al

(2006) The use of speci¢c antibodies and the DIG

probe will enable in exploring cellular localization of

viral proteins In the present study, the polyclonal

antibodies raised against puri¢ed MrNV capsid

pro-tein was successful in locating the viral inclusion

bodies in MrNV-infected muscle tissues by

immunos-taining The viral antigens were detected using

HRP-conjugated secondary antibody and 3-30

diamino-benzidine tetrahydrochloride as substrate The

posi-tive staining in infected prawns con¢rmed the

presence of the virus and the speci¢city of the

poly-clonal antibody produced against MrNV capsid

pro-tein (Fig.7)

The PCR products of MrNV B2 (402 bp) and capsid

(1116 bp) genes were sequenced and the sequence

information of these genes has been deposited in

the GenBank with accession no.s GU300103 and

GU300102 respectively The nucleotide and amino

acid sequence of RdRp for MrNV Indian isolate

(Gen-Bank accession no DQ146969) were compared withTaiwanese (GenBank accession no DQ459205) andChinese (GenBank accession no AY231436) MrNVisolates The nucleotide and amino acid sequence ofMrNV capsid gene were compared with Thai (Gen-Bank accession no EU150126) and PvNV Belize (Gen-Bank accession no EF137180) isolates in addition toTaiwanese (GenBank accession no DQ521575) andChinese (GenBank accession no AY231437) MrNVisolates Presently, the sequence information forRdRp gene of MrNV Thai and PvNV Belize isolatesare not available in the GenBank database

Currently, two genera have been identi¢ed in thefamily Nodaviridae, the alphanodaviruses infectinginsects and betanodaviruses infecting ¢sh (Ball,Hendry, Johnson, Rueckert & Scotti 2000) SriWidada et al (2003) reported complete sequence

of RNA-1 of 3202 bp (Gen Bank accession no.AY222839) encoding an open reading frame (ORF)

of 1045 amino acids for RdRp protein Comparison

of the RdRp gene sequence of MrNV indicated itsproximity to alphanodaviruses (Bonami et al 2005)

In the present study, the partial nucleotide sequence

of the RdRp revealed that the Indian isolate of MrNV(859 bp) did not di¡er much in nucleotide identityfrom the other two isolates of Taiwan (98%) andChina (96%) (Table 1) At the amino acid level, theChinese isolate di¡ered from the Taiwanese and In-dian isolate with respect to one amino acid replace-ment (Arg93Gln) as shown in Fig 8 However, thefull-length gene sequence analysis of RdRp of In-dian isolate of MrNV would be required to know itscomplete sequence homology to other MrNV geo-graphical isolates

The protein sequence analysis of the ORF sequence

of MrNV RdRp used for protein expression (738 bp) ofMrNV (Indian isolate) revealed the signature se-quence along with the conserved core residues,which represent the catalytic domain of RdRp(Fig 9) This observation is based on the reports thathave identi¢ed conserved sequence motifs inRdRp sequence (Poch, Sauvaget, Delarue & Tordo1989) These motifs de¢ne the signature sequence

of RdRp as DX3(FYWLCA)X0 1DXn(STM)GX3TX3(NE)Xn(GS)DD (Koonin & Dolja 1993) It is suggestedthat the presence of signature sequence identi¢es theprotein as a catalytic subunit of viral RNA replicaseand de¢nes it as heart of the polymerase domain(Johnson, Johnson, Dasgupta, Gratsch, & Ball 2001).The motifs of the signature sequence are reported to

be well conserved in RdRp sequences of both insectand ¢sh nodaviruses, even though the overall amino

Figure 7 Immunostaining of MrNV-infected muscles of

Macrobrachium rosenbergii using polyclonal antibodies

raised against MrNV capsid protein ( 400) MrNV,

Macrobrachium rosenbergii nodavirus