Fisheries science JSFS , tập 77, số 3, 2011 5

Growth rates for ages 1–3 of the 1979–1988 year classes, which included low-recruit-ment year classes subsisting during the high population levels of the 1980s, were apparently slower th

O B I T U A R Y

In Memoriam: Professor Michizo Suyama (1923–2011)

Ó The Japanese Society of Fisheries Science 2011

Professor Michizo Suyama, an Honorary Member of the

Japanese Society of Fisheries Science and Former

Profes-sor of Tokyo University of Fisheries, passed away on

January 19, 2011 He was 87 years old

Professor Suyama was born in April 1923 in Tokyo He

graduated from the Imperial Fisheries Institute in 1943 and

became a Research Associate at the Imperial Fisheries

Institute in 1945 In 1949, the National School Establishment

Law created the Tokyo University of Fisheries by

incorpo-rating the Imperial Fisheries Institute and the Faculty of

Fisheries This new institute was placed under the

jurisdic-tion of the Ministry of Agriculture and Forestry Professor

Suyama remained at the new Tokyo University of Fisheries,

becoming an Associate Professor in 1960 and a Full

Pro-fessor in 1974 He retired from the university in 1987

The focus of Professor Suyama’s research activitieswas marine food chemistry His various fields of researchand educational interest included studies on the aminoacid composition of fish protein, extractive components,taste-active components, and smells of volatile substances.His contributions to these fields and those on theimprovement and modification of an amino acid analyzerare excellent and have contributed to developments infisheries science

Professor Suyama also contributed to the JapaneseSociety of Fisheries Science, serving first as a member ofthe Board of Directors, then as Vice President and finallyPresident He received the Japanese Society of FisheriesScience Award of Merit in the fields of nitrogenousextractive components from aquatic animals in 1987 Hissocial contributions as a scientist were invaluable

In 1999, he received The Order of the Sacred Treasure,Gold Rays with Neck Ribbon

Professor Suyama educated and inspired many youngpeople with his profound knowledge and warm personality

He was especially fanatical in his desire to improve thetaste and palatability of food

We offer heartfelt our condolences to the family ofProfessor Suyama

Takaaki ShiraiAssociate Professor, Tokyo University of Marine Scienceand Technology

Fish Sci (2011) 77:289

DOI 10.1007/s12562-011-0346-7

O R I G I N A L A R T I C L E Fisheries

Growth and fatness of 1975–2002 year classes of Japanese sardine

in the Pacific waters around northern Japan

Atsushi Kawabata•Hirotsune Yamaguchi •

Seigo Kubota•Masayasu Nakagami

Received: 3 March 2010 / Accepted: 23 December 2010 / Published online: 26 February 2011

Ó The Japanese Society of Fisheries Science 2011

Abstract We examined individual growth and fatness in

the 1975–2002 year classes of Japanese sardine Samples

were collected at the feeding grounds in the Pacific waters

off northern Japan during drastic fluctuations in the

popu-lation in the 1970s to 2000s Growth rates for ages 1–3 of

the 1979–1988 year classes, which included

low-recruit-ment year classes subsisting during the high population

levels of the 1980s, were apparently slower than for other

year classes There was no obvious trend when comparing

year classes, growth during the first year of life (age 0), and

maximum body length (BL) at age C5 The condition

factors (CF, indicating fatness) for adult sardines of BL

C19 cm in the 1979–1983 year classes during the

maxi-mum population level of the mid-1980s were significantly

lower than for other year classes However, there were no

apparent trends in CF variations for small sardines of BL

\19 cm The apparent decreases in growth rate and fatness

were strongly related to the cumulative sum of population

abundance that each year class experienced It is thought

that insufficient food owing to the density-dependent effect

of an abundant population at feeding grounds resulted in a

decrease in the growth rate for small-bodied sardines thatare investing their energy intake in body growth, and adecrease in fatness for large-bodied adults that are accu-mulating fat for the next reproduction

Keywords Age Body length  Condition factor  Densityeffect Growth  Population  Sardinops melanostictus

IntroductionJapanese sardines Sardinops melanostictus spawn in thecoastal waters along the south of Japan, near Kuroshio, inwinter–spring, and migrate northward in summer–autumn

to forage in the Pacific waters [1 4] The population mass has fluctuated drastically in the past, in synchroni-zation with global climate change, and this fluctuation hasimpacted Japanese fisheries and Japanese society [5] TheJapanese sardine population in the Pacific waters increased

bio-in the 1970s due to the occurrence of comparatively highrecruitment levels in 1973, 1977 and 1978 They becamevery abundant in the 1980s due to very high sequentialrecruitment from 1980 to 1987 (Fig.1) These populationincreases were thought to be due to favorable environ-mental conditions for reproduction and an increase in thenumber of spawned eggs The increased number of eggswas due to the accumulation of spawners from severalgenerations due to the comparatively long life span of thisspecies In the 1980s, over one million tons of sardine werefished, mainly by purse seiner, in the feeding grounds offSanriku and the eastern Hokkaido regions of northernJapan However, the population declined through the1990s, reaching a very low level in the 2000s The sardinecatch in the Pacific waters off northern Japan decreased toonly 0–4000 tons in the 2000s

National Research Institute of Fisheries Science,

Fisheries Research Agency, Japan, 2-12-4 Fukuura,

Kanazawa, Yokohama, Kanagawa 236-8648, Japan

e-mail: abata@affrc.go.jp

H Yamaguchi

Japan Fisheries Information Service Center,

4-5 Toyomi-cho, Chuo-ku, Tokyo 104-0055, Japan

Tohoku National Fisheries Research Institute,

Fisheries Research Agency, Japan, 25-259 Shimomekurakubo,

Samemachi, Hachinohe, Aomori 031-0841, Japan

DOI 10.1007/s12562-011-0326-y

It is known that the individual growth rates of Japanese

sardine in the coastal waters off eastern Hokkaido

decreased in the 1980s compared with the 1970s [8 10] A

similar phenomenon was observed at the past population

peak in the 1930s [11,12] This slow growth was

consid-ered to be subject to a density-dependent effect of the

abundant population [8 10] However, in these studies,

sampling was restricted to the coastal waters of eastern

Hokkaido and so they did not examine differences in

growth rates among year classes There has not been

enough discussion about whether the growth rate in a year

class is affected by its own year class recruitment

abun-dance or rather by the population abunabun-dance (standing

stock of all ages) in Japanese Pacific waters The growth

rate of sardines recovered after the 1990s, but recent

changes in growth related to population abundance have

not yet been examined in depth

Variations in fatness relating to both growth and

popu-lation abundance have also not been examined in detail in

previous studies It was reported that fatness in the 1980s

year classes was lower than in the 1970s year classes when

comparing same-age fish [10] However, there is a positive

correlation between fatness and body length, as described

later in this study The decrease in the fatness of same-age

fish in the 1980s probably reflected shorter body lengths,

rather than fatness per se The effect of fish density on

fatness should be examined clearly by comparing fish of

the same body size

In this study, we examine the growth and fatness of

1975–2002 sardine year classes in relation to the drastic

fluctuations in the population abundance of sardines in the

Japanese Pacific waters, based on biometric data collected

over 31 years We also discuss the density-dependent

effects of recruitment abundance or population abundance

on variations in body length, growth rate and fatnessamong year classes

Materials and methodsFrom 1975 to 2005, we collected samples of Japanesesardine during the fishing season from commercial landings

at the Hachinohe Port, an important fishing port in northernJapan (Fig.2) These were caught by purse seiner off theSanriku coast The fishing season off the Sanriku coastvaried with sardine population abundance, extending to

10 months (from late April to early February) during the1980s due to the high population levels at that time, butshortening to only a few summer months in recent yearsdue to low population levels Vessel survey samples taken

by drift net or pelagic trawl were also collected off thePacific coast of northern Japan in May–January from 1998

to 2003 in order to supplement the decreasing commercialcatch sample (Fig.2)

We measured scaled body length [BL (cm)] and mined age in years (t) by the method of counting annualrings on the scales [16] for a total of 12,161 specimens.These specimens were assigned to the 1975–2002 yearclasses We measured body weight [BW (g)] and gonadweight [GW (g)] for a subset of 11,918 specimens Internalabdominal fat [FW (g)] was also weighed for a subset of7,633 specimens The condition factor (=coefficient offatness, CF) was calculated from the obtained biometricdata using the equation CF = 103(BW – GW)/BL3 Fatweight index was also calculated (=105FW/BL3) Havingexamined the correlation between BL and CF and also theseasonal variation of CF, as described later, we used themean CF for separate BL ranges of samples during May–November (BL \19 cm) or May–December (BL C19 cm)

deter-to examine the effect of density

To estimate the BL at the start of each age, the vonBertalanffy growth curve was fitted by the method ofleast squares to the age and BL data of each year class:

Lt = L? [1 - exp(-K (t - t0))], where Ltis the BL at thestart of age t, L? is the asymptotic BL, K is the rate atwhich BL tends toward the asymptote, and t0 is the agewhen BL = 0 The beginning of each age, the birthday,was set to 1st March, as this was midway through the mainspawning season The age data used to fit the curve was

t ? d/365, where d is the number of days from 1st March

to the sampling date In fitting the curve, L?was limited toless than 30 cm, which is the length of the largest sardinethat has been caught [17]

For the catch of Japanese sardines in the Pacific watersoff northern Japan, we referred to the fishery statisticscollected by the Fisheries Research Agency and to previousreports [18–20], and used them as an index of abundance at

Year

1977, gray and unfilled bars) and recruitment (age 0, gray bars) of

eastern Hokkaido areas of northern Japan (closed diamonds and solid

line)

the feeding grounds For the total population abundance

(standing stock of all ages) and the recruitment of each year

class (cohort abundance at age 0), we referred to the

esti-mates gained from the virtual population analysis

per-formed in Wada and Jacobson [6] and Nishida et al [7]

We examined variations in both the Ltat each age and the

CF for different BLs with respect to population abundance

and year-class recruitment among the 1975–2002 year

classes To examine the effect of population abundance on

the growth of a year class, we used the cumulative sum of

population abundance for the period from age 0 to the

previous age (t - 1) as an index of the abundance that the

year class born in year x experienced:Pxþt1

i¼x Ni; where Ni

is the total population abundance in year i Population

abundance estimates were available for years after 1977

(Fig.1), so year classes after 1977 were examined

Results

Growth

The BLs and ages of the specimens ranged from 6.4 to

24.4 cm and from 0 to 9 years, respectively Figure3

shows age and BL observations and the fitted von

Berta-lanffy growth curves for typical year classes The number

of observations for each year class depended on the ease of

sampling (i.e., year-class abundance) and ranged from 53

(1990 year class) to 2,332 (1986 year class)

Figure4 shows the estimated BL of each year class atthe start of each age from 1 to 9, as estimated from thefitted growth curves Gaps are due to a lack of observationsfor some ages in some year classes The maximum BL atage 5 or older was about 21–22 cm for all year classes,although there was a lack of observations for some yearclasses Estimated BL at age 1 for each year class was13–16 cm Small BLs (13–14 cm) at age 1 occurred inhigh-recruitment (more than 150 billion individuals, asoccurred in 1980–1987) year classes during high popula-tion levels, and in low-recruitment (less than 50 billionindividuals, as occurred in 1979, 1998, 1999 and 2001)year classes during low population levels (Fig 1) Also,large BLs (15–16 cm) at age 1 occurred in both the com-paratively high recruitment (94 billion individuals) yearclass of 1978 and the low-recruitment year classes in the1990s–2000s There was no obvious relation betweenestimated BL at age 1 and either recruitment of year classes

or total population abundance

Estimated BLs at ages 2–4 differed distinctly betweenyear classes Estimated BLs of the 1979–1987 year classeswere 15–16 cm at age 2, 17–18 cm at age 3, and 18–19 cm

at age 4 (Fig.4) These BLs at ages 2–4 were smaller thanthose of the 1975–78 year classes by 1–2 cm and smallerthan those of 1989–2002 year classes by 2–3 cm

Table1 shows the estimated BLs, the growth rates atage t (increase in BL (Lt?1- Lt)/Lt) and the fitted vonBertalanffy equation parameters for typical year classes livingduring different population abundance levels Growth rates

35 40

°N 45

160°E

°N 50

40

30

sampling area, and schematic

view of the oceanographic

current systems and the

spawning ground of the

Japanese sardine population in

Pacific waters (after Watanabe

Cross and diamond symbols

indicate the locations of purse

seiner sampling (n = 340) and

survey vessel sampling

(n = 126), respectively

during age 1 in the 1979 and 1982 year classes (i.e., for

high population levels of more than 300 billion individuals)

were apparently slower than those for the 1987–1989 year

classes (i.e., when the population was decreasing) and for

the 1995 year class (i.e., for low population levels of less

than 30 billion individuals) (Fig.1) However, the

esti-mated BLs at age 1 for the 1979 and 1982 year classes

were not very different to those for other year classes The

growth rates during the early ages for the year classes

corresponding to high population levels (e.g., 1979 and

1982) were slower than those for year classes

corre-sponding to decreasing or low population levels, as

indi-cated by the comparatively small K parameters

Furthermore, the estimated BL at age 2 for the

1988 year class was smaller than those for the 1975–1978

and 1989–2002 year classes (Fig.4; Table1) The growth

rate during age 1 of the 1988 year class should have been

slow, as it is for the 1979–1987 year classes, although there

was a lack of age-1 BL observations However, the

esti-mated BLs at ages 3 and 4 of the 1988 year class were

similar to those for the 1975–1978 and 1989–2002 year

classes (Fig.4; Table1) The BL of the 1988 year class

improved with age The 1987 year class showed a similar

improvement in BL with age (Fig.4; Table 1)

Recruitment in these slow-growing year classes was not

high Recruitment was high (more than 150 billion

indi-viduals) in the 1980–1987 year classes, but was low (less

than 50 billion individuals) in the 1979 and 1988 year

classes (Fig.1; Table 1) Significant correlations were

found between the recruitment and the estimated BLs at

ages 2–4 for the year classes (P \ 0.01) However, there

were some outliers, such as the 1979 year class (Fig.5)

The 1979 year class subsisted during the high-level

popu-lations of the 1980s

On the other hand, sardine populations in the Japanese

Pacific waters were at very high levels during 1980–1989

(306–554 billion individuals) The 1979–1986 year classeswere at ages 1–3 at this time, and had small BLs at ages2–4 (Figs.1,4) Figure6shows the relationships betweenthe estimated BL of each year class at ages 1–4 and thecumulative sum of the population abundance (as explained

in ‘‘Materials and methods’’) in Japanese Pacific waters.Significant negative correlations were noted between the

BL of each age group from 2 to 4 and the cumulative sum

of population abundance, though the BL at age 1 was notsignificantly correlated This suggests that the total popu-lation abundance affected the growth rates during ages 1–3

To examine the effects of the recruitment and totalpopulation abundance on the growth of sardines, we con-ducted multiple regression analysis for the estimated BL atages 2–4 for the year classes against recruitment and thecumulative sum of population abundance (Table2) ForBLs at all ages from 2 to 4, the population abundanceexhibited a large absolute value for the partial regressioncoefficient and was a significant explanatory variable,whereas the recruitment was not a significant variable It

n = 271 1988

n = 1,728 1987

n = 210 1982

n = 169 1979

0 5 10 15 20 25 Body length (cm)

0 1 2 3 4 5 6 7 8 9 10 0

5 10 15 20 25

0 1 2 3 4 5 6 7 8 9 10

Age

0 1 2 3 4 5 6 7 8 9 10

n = 559 1989

Observation Growth curve

age and body length as well as

fitted von Bertalanffy growth

curves for typical year classes

12 14 16 18 20 22 24

1975–2002 year classes

was the total population abundance rather than the

recruitment that strongly affected the BLs of sardines at

ages 2–4

Fatness

Almost all specimens, including adults near the maximum

BL, had small gonads Most specimens had large amounts

of subcutaneous and internal abdominal fat Fat weight

index was positively correlated to the CF (r = 0.560,

n = 7,633, P \ 0.01) The CFs of all specimens were in

the range 7.9–17.2, with an average of 12.6 There was a

weak but significant trend for sardines with a larger BL to

have a higher CF (r = 0.395, n = 11,918, P \ 0.01) A

comparison of the CF at the same age between year classes

could reflect the differences in BL mentioned above

Therefore, the variation in CF between year classes wasexamined for separate BL ranges Also, to examine sea-sonal variations in CF, monthly mean CFs for each yearclass was determined The means for all year classes werethen averaged (Fig.7) The mean CF in each month can beseen to decrease in winter: December–February for BLs

\19 cm or in January for BLs C19 cm The foragingseason for Japanese sardine in the Pacific waters is summerand autumn [21] To examine the variations in CF amongyear classes, we used the comparatively high CF valuesobtained from samples during May–November for BL

\19 cm or during May–December for BL C19 cm as thevalues that were considered to occur in the foraging season.The mean CFs by BL range in each year class are shown

in Fig 8 The mean CFs for BL C19 cm were about 13–14

Body length (cm): L t

1979 1988

beginning of each age (t) and recruitment for the 1975–2002 year

22 21 20

19 18

20 19 18 17 16 15 14

Cumulative sum of population abundance (109ind.)

1000

beginning of each age (t) for the 1977–2002 year classes and the cumulative sum of population abundance for the period from age 0 to age (t - 1) (**P \ 0.01)

for most of the year classes, but were significantly lower

(11.7–12.4) for the 1979–1983 year classes The mean CFs

of each year class for BL = 17–19 cm and BL = 15–17 cm

were about 12–13 and about 11–12, respectively, and those for

BL \15 cm ranged from 10 to 12 There were no apparent

trends in CF variation for BL \19 cm, which was similar to

what was seen for BL C19 cm

The mean CFs of large-bodied sardines with BL

C19 cm during each year are shown along with the catch in

Fig.9 The CF values in 1983–1988 were in the range

11.4–12.2, lower than in most other years During this time,

population levels in the Japanese Pacific waters were high

(Fig.1), and the catch levels around northern Japan (an

index of abundance in the feeding grounds) were at a

maximum There was a significant negative correlation

between CF and the catch (r = 0.578, n = 26, P \ 0.01)

These large-bodied sardines of low CF mainly consisted of

1979–1983 year classes The high abundance in the

feed-ing grounds must have affected the fatness of large-bodied

sardines

Discussion

We compare our results with growth patterns for other

small pelagic fishes that show biomass fluctuations like

Japanese sardine Previous studies have reported

density-dependent growth for some small pelagic fishes The

western North Pacific chub mackerel (Scomber japonicus)

undergoes a seasonal migration similar to that of Japanese

sardine Their body lengths at age C1 are dependent on

their growth in age 0 and are negatively correlated with the

year-class strength [22] The body size (length, mass) at

age 3 of the Pacific herring (Clupea pallasii) from the

southwest coast of Vancouver Island is negatively related

to parental biomass, owing to a pre-recruit effect at age 0

[23] The body length at age of Pacific Hokkaido spring

spawning herring is highly dependent on growth during age

0, and exhibited weak density-dependent effects of

year-class strength [24] Thus, the body sizes at older ages in

these species are determined by their growth early in life,

which is in turn affected by own recruitment level orparental biomass (i.e., egg abundance)

Our results indicate different growth patterns for nese sardine from these species Year classes subsistingduring high population levels showed apparent decreases inthe growth rate of small-bodied sardines and in the fatness

Japa-of large-bodied sardines (Figs.4,8) These decreases weremore strongly related to the cumulative sum of populationabundance that each year class experienced, rather than therecruitment level (Fig.6; Table2) Significant correlationsthat were found between the recruitment and the estimatedBLs were unlikely to indicate a direct effect of recruitmentabundance on the growth (Fig.5)

Density-dependent growth was not apparent during thefirst year of life (age 0); there were no obvious relation-ships between estimated BL at the beginning of age 1 andrecruitment or total population level (Figs.5, 6) Insuffi-cient food due to the density effect during age 0 would not

be serious compared with age C1 Juvenile sardines of age

0 have less swimming ability and are thought to beentrained passively by ocean currents such as the KuroshioExtension During their denatant feeding migration, theyare distributed broadly from the coastal area to around theShatsky Rise and far off the Kuril Islands to about 165°E inthe Pacific Ocean This is regardless of the population level

or the recruitment level [25–27] Age 0 sardines are welldispersed and able to use the broad waters as foraging areas

Standard partial regression coefficient

May Jun Jul Aug Sep Oct Nov Dec Jan Feb

Mean CF

BL

>19 cm 17-19 cm 15-17 cm

<15 cm

(even if they are not in the best habitat), so an obvious

density effect would not occur After wintering, age C1

sardines actively migrate, looking for good feeding

grounds to forage in Consequently, the sardines compete

intra-specifically in favorable feeding grounds and have to

migrate further to find more food In such a situation, the

high population abundance caused an obvious density

effect It is suggested that decreases in growth rates or

fatness were caused by a decrease in energy intake due to

high density at the feeding grounds

Lower CF was observed in large-bodied sardines (BL

C19 cm) during the period of maximum abundance at the

feeding grounds in the mid-1980s (Figs.8,9)

Large-bod-ied sardines of BL C19 cm are near to the maximum body

size for this species and have a slow growth rate in terms of

body length They are adults and accumulate the food

energy taken in at the feeding ground during

summer-autumn as body fat for reproduction in the following

winter-spring [21] Small-bodied sardines of BL \19 cm

have higher growth rates than large-bodied sardines and

would invest their energy intake into increasing their body

lengths During the period of maximum population, the

effect of high density on small sardines emerged as a

decrease in growth rate for 10 year classes (1979–1988)

over a comparatively long period: the effect on

large-bodied sardines appeared as a decrease in fatness for only

5 year classes (1979–1983) It seems that small-bodiedsardines were more affected by the density than larger-bodied sardines The mean BL of sardines caught duringthe 1980s in the waters off eastern Hokkaido (further fromthe spawning ground) was greater than that of sardinescaught in the coastal waters off Sanriku [20] This maymean that larger-bodied sardines can forage farther afieldthan smaller sardines and so reduce the density effect

11 12 13

<15 cm

9 10 11 12 13

12 13 14

range for the 1975–2002 year

classes

Year

CF of BL >19 cm Catch in numbers SD

11 12 13 14 CF

0 5 10 15 20 25 30 Catch in numbers (x10 9 ind.)

in the Pacific waters off Sanriku and the eastern Hokkaido areas of

According to previous studies based on analyses of

ancient documents relating to Japanese regional fisheries,

extra big catch years due to explosive increases in the

Japanese sardine population in Pacific waters have

occur-red only five or six times since the sixteenth century [3,11,

28] These previous periods of population increase seemed

to last no longer than a few decades The recent explosive

population increase in the 1980s was considered to have

lasted for about a decade Only 8 year classes (1980–1987)

had outbreaks contributing to the explosive increase in

population during the 1980s (Fig.1) This explosive

pop-ulation increase (outbreak) caused an obvious density

effect, as mentioned above Changes in biological

charac-teristics owing to the density effect, such as an apparent

decrease in growth rate or fatness, were conspicuous and

seemed to be unusual for Japanese sardine Additionally,

the condition of spawning sardines (i.e., accumulation of

body fat during the previous summer’s feeding migration)

affects their gonadal development and the quality and

quantity of eggs [29–31] Kawasaki and Omori [32] state

that the density effect at the feeding ground in the 1980s

caused the condition of the spawners and the egg quality at

the spawning ground to deteriorate The outbreak was not

necessarily a favorable situation for the sardine population

During the long intervals between outbreaks of the

popu-lation, the year classes that occur show relatively low

recruitment and biological characteristics that might be

regarded as normal The outbreak of Japanese sardines was

a phase variation that coincided with global climate change

[33] This does not indicate a recovery of the sardine

population; instead it may be a short-term, disadvantageous

state for the population

two anonymous reviewers for their constructive comments and

suggestions We also thank Ms J Momosawa, Ms N Kubo,

Mr M Kawamura and other staff at the Tohoku National Fisheries

Research Institute for their assistance and support in this study This

study was funded by the Program of Marine Fisheries Stock

Assessment and Evaluation for Japanese Waters from the Fisheries

Agency of Japan This paper is a contribution to the Study for the

Prediction and Control of Population Outbreak in Marine Life in

Relation to Environmental Change (POMAL) of the Agriculture,

Forestry and Fisheries Research Council (AFFRC).

References

1 Kondo K (1980) The recovery of the Japanese sardine—the

biological basis of stock-size fluctuations Rapp P-v Re´un Cons

Int Explor Mer 177:332–354

2 Kondo K (1988) General trends of neritic-pelagic fish

popula-tions—a study of the relationships between long-term fluctuations

of the Japanese sardine and oceanographic conditions (in

Japa-nese) In: Japanese Society of Fisheries Oceanography (ed)

Studies on fisheries oceanography Koseisha Koseikaku, Tokyo,

5 Kawasaki T (1992) Climate-dependent fluctuations in far eastern sardine population and their impacts on fisheries and society In: Glantz MH (ed) Climate variability, climate change and fisheries Cambridge University Press, Cambridge, pp 325–354

6 Wada T, Jacobson LD (1998) Regimes and stock-recruitment relationships in Japanese sardine (Sardinops melanostictus), 1951–1995 Can J Fish Aquat Sci 55:2455–2463

7 Nishida H, Ishida M, Kawabata A, Watanabe C (2008) Stock assessment of the Japanese sardine in fiscal 2008 year (in Japa- nese) In: Fisheries Research Agency (ed) Marine stock assess- ment and evaluation for Japanese waters, Japan Fisheries Agency, Tokyo, pp 11–42

8 Mihara Y (1988) The sardine fisheries in the waters off eastern Hokkaido (in Japanese) Hokusuishi Dayori (4):1–6

south-9 Wada T, Kashiwai M (1991) Changes in growth and feeding ground of Japanese sardine with fluctuation in stock abundance In: Kawasaki T et al (eds) Long-term variability of pelagic fish populations and their environment Pergamon, London,

pp 181–190

10 Wada T (1998) Migration range and growth rate in the Oyashio area (in Japanese) In: Watanabe Y et al (eds) Stock size fluctu- ations and ecological changes of the Japanese sardine Koseisha Koseikaku, Tokyo, pp 27–34

11 Ito S (1961) Fishery biology of the sardine, Sardinops anosticta (T & S.), in the waters around Japan (in Japanese with English abstract) Bull Jap Sea Reg Fish Res Lab (9):1–227

mel-12 Nakai Z (1962) Studies relevant to mechanisms underlying the fluctuation in the catch of the Japanese sardine, Sardinops mel- anosticta (Temminck & Schlegel) Jap J Ichthyol 9:1–115

13 Watanabe T (1983) Stock assessment of common mackerel and Japanese sardine along the Pacific coast of Japan by spawning survey (FAO Fisheries Report 291) (2):57–81

14 Kuroda K (1988) Yearly changes of the main spawning grounds

of the sardine, Sardinops melanostictus (T et S.) in the waters along the Pacific coast of southern Japan (in Japanese with English abstract) Bull Jpn Soc Fish Oceanogr 52:289–296

15 Kobayashi M, Kuroda K (1991) Estimation of main spawning grounds of the Japanese sardine from a viewpoint of transport condition of its eggs and larvae In: Kawasaki T et al (eds) Long- term variability of pelagic fish populations and their environment Pergamon, London, pp 109–116

16 Kondo K (1964) Life history of Japanese sardine, Sardinops melanosticta (Temminck and Schelegel), and a proposal meth- odology on the investigations (in Japanese) Japan Fisheries Resource Conservation Association, Tokyo

17 Adachi J (1989) The big sardine caught off Hamada, Japan (in Japanese) Bull Jpn Soc Fish Oceanogr 53:334–335

18 Murakami K, Kobayashi T (1978) Surveys on coastal important fisheries resources (in Japanese) Annu Rep Hokkaido Kushiro Fish Exp Stn 83–92

19 Wada T (1988) Population dynamics on Japanese sardine, Sardinops melanostictus, caught by the domestic purse seine fishery in the waters off the coast of southeastern Hokkaido (in Japanese with English abstract) Bull Hokkaido Reg Fish Res Lab (52):1–138

20 Yamaguchi H, Kawabata A (1992) Characteristics of the purse seiner’s fishing condition and the available year-classes of Jap- anese sardine, Sardinops melanostictus, in the waters off the coast

of northern Sanriku, 1984–1990 (in Japanese with English abstract) Bull Tohoku Natl Res Inst (54):23–58

21 Kondo K, Hori Y, Hiramoto K (1976) Life pattern of the Japanese sardine, Sardinops melanosticta (Temmink et Schlegel), and its

practical procedure of marine resources researches of the stock

(2nd edn) (in Japanese) Japan Fisheries Resource Conservation

Association, Tokyo

22 Watanabe C, Yatsu A (2004) Effects of density-dependence and

sea surface temperature on interannual variation in length-at-age

of chub mackerel (Scomber japonicus) in the Kuroshio-Oyashio

area during 1970–1997 Fish Bull 102:196–206

23 Tanasichuk RW (1997) Influence of biomass and ocean climate on

the growth of Pacific herring (Clupea pallasi) from the southwest

coast of Vancouver Island Can J Fish Aquat Sci 54:2782–2788

24 Watanabe Y, Dingsor GE, Tian Y, Tanaka I, Stenseth NC (2008)

Determinants of mean length at age of spring spawning herring

off the coast of Hokkaido, Japan Mar Ecol Prog Ser 366:209–217

25 Kenya VS (1982) New data on the migrations and distribution of

Pacific sardines in the Northwest Pacific Sov J Mar Biol 8:41–48

26 Sugisaki H (1996) Distribution of larval and juvenile Japanese

sardine (Sardinops melanostictus) in the western North Pacific

and its relevance to predation on these stages In: Watanabe Y

et al (eds) Survival strategies in early life stages of marine

resources AA Balkema, Rotterdam, pp 261–270

27 Kinoshita T (1998) Northward migration juveniles in the

Kuro-shio Extension area (in Japanese) In: Watanabe Y et al (eds)

Stock size fluctuations and ecological changes of the Japanese

sardine Koseisha Koseikaku, Tokyo, pp 84–92

28 Tsuboi M (1988) Main spawning ground of Japanese sardine around Japan (in Japanese) Tokai Reg Fisheries Res Lab Sakana 40:37–49

29 Morimoto H (1996) Effects of maternal nutritional conditions on number, size and lipid content of hydrated eggs in the Japanese sardine from Tosa Bay, southwestern Japan In: Watanabe Y et al (eds) Survival strategies in early life stages of marine resources.

AA Balkema, Rotterdam, pp 3–12

30 Morimoto H (1998) Maturation (in Japanese) In: Watanabe Y

et al (eds) Stock size fluctuations and ecological changes of the Japanese sardine Koseisha Koseikaku, Tokyo, pp 45–53

31 Shiraishi M, Ikeda K, Akiyama T (1996) Effects of water perature and feeding rate on gonadal development in the Japanese sardine (Sardinops melanostictus) in captivity In: Watanabe Y

tem-et al (eds) Survival strategies in early life stages of marine resources AA Balkema, Rotterdam, pp 13–19

32 Kawasaki T, Omori M (1995) Possible mechanisms underlying fluctuations in the Far Eastern sardine population inferred from time series of two biological traits Fish Oceanogr 4:238–242

33 Kawasaki T (1991) Long-term variability in the pelagic fish populations In: Kawasaki T et al (eds) Long-term variability of pelagic fish populations and their environment Pergamon, London,

pp 47–60

O R I G I N A L A R T I C L E Biology

Expression patterns of type II and III iodothyronine deiodinase

genes in the liver of the goldlined spinefoot, Siganus guttatus

Nina Wambiji•Yong-Ju Park• Ji-Gweon Park•

Se-Jae Kim•Sung-Pyo Hur•Yuki Takeuchi•

Akihiro Takemura

Received: 17 November 2010 / Accepted: 23 January 2011 / Published online: 10 March 2011

Ó The Japanese Society of Fisheries Science 2011

Abstract Iodothyronine deiodinases play an important

role in thyroid hormone regulation in vertebrates The aim

of this study was to clone type II (SgD2) and type III

(SgD3) iodothyronine deiodinase cDNA from the goldlined

spinefoot (Siganus guttatus) using 30- and 50-rapid

ampli-fication of cDNA ends and then to assess their expression

patterns in the liver under several experimental conditions

by using quantitative real-time PCR SgD2 (1013 bp) and

SgD3 (1492 bp) contained open reading frames of 810 and

804 bp and encoded 270 and 269 amino acids,

respec-tively They were characterized by an in-frame TGA codon

that was considered to be selenocysteine An abundance of

SgD2 and SgD3 mRNA was expressed in several tissues,

with an increase at 1200 hours and a decrease at

2400 hours Food deprivation suppressed the expression of

SgD2, but not SgD3 Higher SgD2 and SgD3 mRNA levels

in the liver were found in fish reared at 25°C than in those

reared at 20 and 30°C These results suggest that

exoge-nous factors influence the mRNA levels of iodothyronine

deiodinase genes in the liver and that transcription of the

genes in certain tissues is partially regulated in a circadianmanner

Keywords Cloning Day–night variations  Foodavailability Rabbitfish  Quantitative real-time PCR Temperature

IntroductionThere are two types of thyroid hormones (THs), namely,3,5,30,50-tetraiodothyronine (T4) and 3,5,30-triiodothyro-nine (T3), both of which play important roles in the phys-iological aspects of growth, development, and reproduction[1,2] T3is the potent and biologically active form of THand is produced by the enzymatic outer-ring deiodination(ORD) of T4 in extrathyroidal tissues In contrast, thegeneration of 3,30,50-triiodothyronine (reverse T3 or rT3),the inactive form of TH, is produced by inner-ring deio-dination (IRD) [3 5] ORD and IRD are also active in themetabolic pathways that form the inactive compound 3,30-diiodothyronine (T2) from T3and rT3, respectively [2] Thedeiodination processes that occur during ORD and IRD areconsidered to be tissue specific and to regulate intracellular

TH availability and disposal [6] Iodothyronine ases, which are members of the selenoprotein family, arethe enzymes responsible for TH deiodination [4,7] Threetypes of iodothyronine deiodinases have been identified invertebrates [7]: type-I (D1) enzymes possess ORD and IRDactivities, while type-II (D2) and type-III (D3) only haveORD and IRD activity, respectively [4,8] It appears thatthe expression patterns of iodothyronine deiodinases inrespective organs are species specific and can vary in thesame species depending on the organism’s physiologicalstatus [9]

Department of Chemistry, Biology and Marine Sciences,

Faculty of Science, University of the Ryukyus, Nishihara,

Okinawa 903-0213, Japan

e-mail: takemura@sci.u-ryukyu.ac.jp

Department of Biology, Jeju National University,

66 Jejudaehakno, Jeju City, Jeju Special Self-Governing

Province 690-756, Republic of Korea

Y.-J Park

Marine and Environmental Research Institute, Jeju National

University, 3288 Hamduk, Jocheon, Jeju Special Self-Governing

Province 695-814, Republic of Korea

Fish Sci (2011) 77:301–311

DOI 10.1007/s12562-011-0330-2

In certain teleost fishes, deiodination activities of TH

occur primarily in the liver [10,11] A deiodination assay

using radiolabeled iodine demonstrated high levels of

activity of low-KmT4ORD (the functional equivalent of D2)

in the liver of a large number of teleost fishes [6,11–15]

Plasma T3levels have been found to be highly correlated to

T4ORD activity in the liver of the Atlantic salmon Salmo

salar, suggesting that this organ is a major source of

cir-culating T3 in teleosts [16] On the other hand, D3-like

activity has also been reported in the liver of salmonids

[17, 18], sturgeon Acipenser fulvescens [19], walleye

Sander vitreus [20], American plaice Hippoglossoides

platessoides [21], and Nile tilapia Oreochromis niloticus

[22] Feeding Nile tilapia and rainbow trout Oncorhynchus

mykiss with T3-supplemented food resulted in an increase in

D3 activity in the liver and gills but not in the brain and

kidney [23–25], while rainbow trout immersed in a solution

supplemented with T4showed induced D3 activity in the

brain, liver, and retina [26] These findings for D2 and D3

indicate that alternations in hepatic iodothyronine

deiodin-ase activity impact on the TH-bdeiodin-ased status in certain

peripheral organs

Acclimation to low temperature conditions was

observed to decrease D2 activity in the liver of the Atlantic

cod Gadus morhua [14] The enzymatic activities of D2

and D3 in fish are also affected by nutritive and stress

conditions [27,28], and D2 activity in the liver has been

shown to respond to sex steroids and pituitary hormones

[29–31] These results demonstrate that the hepatic

deio-dination processes are directly or indirectly affected by

endogenous and exogenous factors [2] To date, most

studies on THs in fish have focused on the effects of

environmental factors on deiodinase enzyme activities, and

few studies have employed molecular approaches to

eval-uate the effects of such factors on the expression of

iodo-thyronine deiodinase genes in fish, although those genes

have been fully cloned and characterized in Nile tilapia [3]

and rainbow trout [32] The aim of the study reported here

was to assess the molecular characteristics of

iodothyro-nine deiodinases in the liver of the goldlined spinefoot

(formerly referred to as the golden rabbitfish or

orange-spotted spinefoot), Siganus guttatus, a common coral reefs

species and an important fish resource in Southeast Asian

countries Since TH is closely related to reproductive and

nutritive conditions, understanding the status of D2 and D3

in the liver may lead to improved aquaculture technologies

for this species We first cloned the cDNA for D2 and D3

from this species (SgD2 and SgD3, respectively) After

establishing an assay system to measure D2 and D3 mRNA

levels by using quantitative real-time PCR (qPCR), we

assessed day–night differences in SgD2 and SgD3 mRNA

abundance in several tissues, as well as the effects of food

deprivation and temperature on their expression in theliver

Materials and methodsFish

Juvenile goldlined spinefoot with a body mass of0.08–0.15 g were caught using small-mesh nets from themangrove estuary of the Teima River, Northern Okinawa,Japan, during daytime, at low tide, around the time of thenew moon in July and August They were reared undernatural photoperiodic conditions in holding tanks (capac-ity 1 or 5 t) containing constantly flowing seawater atambient temperature at the Sesoko Station, TropicalBiosphere Research Center, University of the Ryukyus,Okinawa, Japan The fish were fed daily at 1000 hourswith commercial pellets (EP1 and then EP2; MarubeniNisshin, Tokyo, Japan) Immature fish with a mean bodymass of 200 ± 0.5 g (age 1) and mature fish with a meanbody mass of 346 ± 0.5 g (age 3 and 4) were used in thestudy

All experiments were conducted in compliance with theAnimal Care and Use Committee guidelines of the Uni-versity of the Ryukyus and with the regulations for the careand use of laboratory animals in Japan

Sample collections for molecular cloning and tissueexpression

The fish were transferred to outdoor polyethylene tanks(capacity 300 l) containing running seawater and accli-mated to the rearing conditions with a fixed food provi-sion at 1000 hours for 1 week The fish were taken fromthe tanks at 1200 hours, anesthetized with 2-phenoxy-ethanol (Kanto Chemical, Tokyo, Japan), and immediatelykilled by decapitation The whole brain was taken fromthe mature fish (n = 3) for the molecular cloning of SgD2and SgD3 cDNA For the analysis of the tissue distribu-tion of SgD2 and SgD3 mRNA, the whole brain, retina,gills, liver, spleen, kidney, gonads, heart, and skin werecollected at 1200 hours (n = 7) and 2400 hours (n = 7)from immature fish that were acclimated under the sameconditions The samples were immediately immersed inRNAlater (Applied Biosystems, Foster City, CA) at 4°Cand then stored at -20°C until further analysis Duringthe dark period, samples were collected under a dim redlight Since the day/night variations of both deiodinasesamong the tissues were higher during the day than atnight, we subsequently collected all our samples at

1200 hours

Effects of food deprivation and temperature on gene

expression

For the food deprivation experiment, immature fish were

transferred to two polyethylene tanks (capacity 300 l)

containing running seawater at 25 ± 1.0°C under natural

photoperiodic conditions The fish (n = 7) in one tank

were fed daily at 1000 hours with commercial pellets at 5%

of their body mass, while the fish (n = 7) in the other tank

were not fed after the initiation of the experiment After

rearing under these conditions for 1 week, the fish from

each tank were anesthetized and sacrificed at 1200 hours

After weighing, blood was collected from the caudal vein

using heparinized syringes and centrifuged at 10,000 g for

10 min at 4°C to obtain plasma samples that were stored at

-20°C until their glucose levels could be measured The

liver was then taken from the body cavity and weighed

Pieces of the liver were immersed in RNAlater at 4°C and

stored at -20°C until analysis The hepatosomatic index

(HSI) was calculated using the following formula:

HSI = (liver mass/body mass) 9 100

For the temperature experiment, immature fish (8 fish

per aquarium) were transferred to three glass aquaria

(capacity 60 l) with running seawater at 25°C, maintained

under LD = 12:12 by placing a fluorescent lamp (20 W)

above each aquarium that provided illumination at 1200 lx

(light intensity 2.23 W/m2), which was measured using a

quantum photoradiometer (model HD 9021, Delta OHM,

Padova, Italy) After acclimation for 1 week, the

temper-ature of each aquarium was gradually changed using a

temperature control system with a programmable set point

to 20°C (lowest in winter), 25°C (temperature during

spawning season), and 30°C (highest in summer) The fish

were fed with commercial pellets daily at 1000 hours at 5%

of their body mass One week after rearing the fish under

these conditions, samples were collected as mentioned

above

Measurement of glucose levels

Plasma glucose levels were determined using the Glucose

CII Test Wako kit (Wako Pure Chemical Industries, Osaka,

Japan) according to the manufacturer’s instructions

Extraction of RNA and cDNA synthesis

Total RNA was extracted from the tissues using the

Tri-Pure Isolation Reagent (Roche Applied Science, Hague

Road, IN) according to the manufacturer’s instructions

When necessary, the samples for qPCR were treated with

deoxyribonuclease (RT grade; Nippon Gene, Tokyo,

Japan) at 37°C for 15 min to avoid contamination with

genomic DNA The amount of RNA was measured at

260/280 nm, and samples with an absorbance ratio (A260/

A280) of 1.8–2.0 were used for cDNA synthesis

Complementary DNA (cDNA) was reverse-transcribedfrom 0.5 lg of total RNA using the High Capacity cDNAReverse Transcription kit (Applied Biosystems) for qPCRand molecular cloning according to the manufacturer’sinstructions For cloning, the first strand cDNA was syn-thesized from 1 lg of total RNA using PrimeScript 1ststrand cDNA Synthesis kit (Takara Bio, Otsu, Japan).Cloning of the iodothyronine deiodinase genesThe SgD2 and SgD3 cDNA fragments were amplifiedusing degenerate oligonucleotide primers (SgD2-F2 andSgD2-R1 for SgD2, and SgD3-F1 and SgD3-R2 for SgD3)that were designed on the basis of the highly conservedregions of the target genes using Primer3 software(Whitehead Institute/Massachusetts Institute of Technol-ogy, Boston, MA) (Table 1) Oligonucleotide primers weredesigned on the basis of the D2 sequences of the tigerpufferfish Takifugu rubripes (GenBank accession no.AB360768), bastard halibut Paralichthys olivaceus(AB362422), and mummichog Fundulus heteroclitus(FHU70869), and on the basis of D3 sequences of the Niletilapia (Y11111), tiger pufferfish (AB360769), and bastardhalibut (AB362423) (Table 1) The PCR was performed in

SgD2 and SgD3 of goldlined spinefoot (Siganus guttatus)

25 ll of sample with Go Taq Green Master Mix (Promega,

Madison, WI) under the following cycling conditions: 1

cycle of initial denaturation for 2 min at 94°C; 35 cycles of

denaturation at 94°C for 45 s, annealing at 58°C for 45 s,

and 72°C for 1 min The PCR products were separated on a

1% agarose gel with an appropriate molecular weight

marker, stained with ethidium bromide, and visualized

under UV illumination (ATTO, Tokyo, Japan) When the

PCR products of the predicted sizes were obtained, these

were purified using the Wizard SV Gel and PCR Clean-up

System kit (Promega) and ligated The purified products

(233 and 281 bp for SgD2 and SgD3, respectively) were

then cloned into the pGEM T-Easy Vector (Promega) and

sequenced

Rapid amplification of cDNA ends (RACE) was carried

out using the SMART RACE cDNA Amplification kit

(Clontech Laboratories, Mountain View, CA) according to

the manufacturer’s instructions On the basis of the

sequence of the partial cDNA fragments described above,

the specific primers and nested primers for the RACE of

SgD2 (SgD2-GSP2, SgD2-NGSP2, SgD2-F1-GSP1, and

SgD2-F3-NGSP1) and those for the RACE of SgD3

(GSP2, NGSP2, GSP1, and

SgD3-NGSP1) were designed for the 50- and 30-ends, respectively

(Table1) RACE reactions in the first PCR were performed

using the Universal Primer A Mix (UPM) and the

gene-specific primer in a three-step touchdown PCR program:

(1) 5 cycles of 94°C for 30 s and 72°C for 3 min; (2) 5

cycles of 94°C for 30 s, 70°C for 30 s, and 72°C for 3 min;

(3) 25 cycles of 94°C for 30 s, 68°C for 30 s, and 72°C for

3 min Nested PCR was performed using the 20-fold

diluted first PCR products as a template with the Nested

Universal Primer A (NUP) and each gene-specific nested

primer at the following cycling conditions: 94°C for 2 min;

25 cycles of 94°C for 30 s, 68°C for 30 s, and 72°C for

2 min; a final step of 72°C for 3 min [33] The cDNA

fragments amplified by RACE were cloned into the

pGEM-T Easy vector and then sequenced

Sequence analysis

The nucleotide and deduced amino acid sequences were

analyzed using the BLAST program (http://ncbi.nlm.nih

gov/BLAST) Multiple alignments for phylogenetic

anal-ysis were performed using the full-length deiodinase

sequences of several vertebrates by the ClustalW program

(http://www.ebi.ac.uk/clustalw)

Real-time quantitative PCR (qPCR)

The expression levels of SgD2 and SgD3 mRNA were

assessed using the CFX96 Real-Time System C1000

thermal cycler (Bio-Rad, Hercules, CA) The forward and

reverse primers for the qPCR (qPCR-F and qPCR-R for SgD2; SgD3-qPCR-F and SgD3-qPCR-R forSgD3; b-actin-qPCR-F and b-actin-qPCR-R for b-actin)were designed as shown in Table1 b-actin mRNA levels

SgD2-in the same sample were determSgD2-ined usSgD2-ing qPCR to malize the expression data [33] The qPCR reaction mix-ture (10 ll) contained 5 ll Express SYBR GreenER qPCRSupermix Universal (Invitrogen, Carlsbad, CA), 0.3 lMforward primer, 0.3 lM reverse primer, 2 ll cDNA tem-plate, and 2.4 ll RNase-free water The following PCRcycling conditions were used: 95°C for 30 s; 40 cycles of95°C for 5 s and 60°C for 34 s To ensure the specificity ofthe PCR amplicons, a melting curve analysis was carriedout by raising the temperature of the sample slowly from

nor-60 to 95°C until the final step of the PCR The expressionlevels of SgD2, SgD3, and b-actin mRNA were measured

in triplicate Data were normalized relative to the meanexpression level of each gene and analyzed using thenormalized gene expression [DDC(t)] method [34].Statistical analysis

All data were expressed as the mean ± standard error ofthe mean (SEM) Normality was tested using the Kol-mogorov–Smirnov method Student’s t test and the Mann–Whitney U test were used to analyze the statistical differ-ences between two sets of data A one-way analysis ofvariance (ANOVA) was performed for the temperatureexperiment Probabilities of P \ 0.05 and P \ 0.01 wereconsidered to be statistically significant

ResultsCloning and properties of SgD2 and SgD3The RACE analyses of SgD2 cDNA yielded a 1013 bpfragment with an open reading frame (ORF) of 810 bp(Fig.1) The predicted amino acid sequence was 270 res-idues long The ORF was interrupted by an in-frame TGAcodon at position 591 that was likely to encode for sele-nocysteine (Sec); however, the sequence did not contain aSECIS element that has been reported to be responsible forthe incorporation of Sec into the protein during translation.The SgD3 cDNA consisted of a 1492 bp fragment with anORF of 804 bp The Sec residue was at position 504 Thepredicted amino acid sequence of SgD3 was 269 residueslong (Fig 2)

In terms of the amino acid sequence, SgD2 showed ahigh similarity with D2 from several teleosts, such as thegilthead seabream Sparus aurata (90%), bastard halibut(85%), medaka Oryzias latipes (83%), and tiger pufferfish(80%) It showed a moderately high similarity with D2

from the rat Rattus norvegicus (69%) and chicken Gallus

gallus (70%) SgD3 also showed a high similarity with D2

from gilthead seabream and Nile tilapia (90%), Senegalese

sole Solea senegalensis (86%), bastard halibut (83%), and

Atlantic halibut Hippoglossus hippoglossus (79%) and had

equally high similarities with D2 from the cow Bos taurus

(67%) and chicken (72%) (data not shown)

Tissue distribution of iodothyronine deiodinase genes

The tissue distribution of SgD2 and SgD3 was examined at

1200 and 2400 hours using qPCR SgD2 and SgD3 mRNA

were detected in all of the tissues tested A comparison of

the expression of SgD2 mRNA in various tissues at

1200 hours revealed high expression in the liver, brain,

skin, and spleen (Fig.3a) High levels of SgD3 mRNA

were observed in the liver, brain, retina, spleen, and skin

(Fig.3b)

Day–night differences in the abundance of SgD2 and

SgD3 mRNA were observed The expression levels of SgD2

in the liver, skin, brain, heart, and gonads were significantlyhigher at 1200 hours than at 2400 hours A significantlyhigher expression of SgD3 mRNA during the daytime wasalso observed in the liver, retina, brain, gills, heart, gonads,and skin Negligible levels of SgD2 mRNA were observed

in the gonads and heart, while the smallest amount of SgD3mRNA was detected in the kidney followed by the gills andthe heart at 2400 hours (Fig.3a, b)

Food deprivationThe effects of food deprivation on HSI and glucose wereexamined at 1200 hours Following food deprivation, HSIand plasma glucose levels significantly decreased (Fig.4a,b) and glycogen levels in the liver dropped (data notshown) The effects of food deprivation on SgD2 and SgD3mRNA expression in the liver at 1200 hours are shown inFig.5 Food deprivation significantly lowered the abun-dance of SgD2 mRNA (P \ 0.05) (Fig.5a), but not ofSgD3 (Fig.5b)

amino acid sequence of the

Siganus guttatus iodothyronine

deiodinase type II (SgD2)

cDNA clone The complete

mRNA spans 1013 bp with an

open reading frame (ORF) of

810 bp (270 amino acids) Sec

in bold Selenocysteine residue

at position 591 The start and

stop codons are indicated in

bold, with the stop codon

denoted by an asterisk

Effects of SgD2 and SgD3 mRNA expression were

examined in the liver for the range of water temperatures

encountered in the habitats of the goldlined spinefoot

(Fig.6) We observed that temperature significantly

affected the expression of SgD2 and SgD3 mRNA

(P \ 0.01) The levels of SgD2 transcription were

signifi-cantly higher (P \ 0.01) at 25°C than at 20 and 30°C

(Fig.6a) A similar temperature effect was observed for

SgD3 mRNA levels (P \ 0.01) (Fig.6b)

Discussion

The first step of this study was the cloning and

character-ization of cDNA encoding type II (SgD2) and type III

(SgD3) iodothyronine deiodinase of the goldlined

spine-foot, with the aim of evaluating the effects of food

depri-vation and temperature on SgD2 and SgD3 mRNA

abundance in the liver The ORF of SgD2 and SgD3

con-tained an in-frame TGA stop codon that is characterized by

the presence of selenocysteine The incorporation of an

essential selenocysteine residue within the catalytic domain

requires the presence of a premature stop codon (TGA) inthe ORF and a SECIS element located in the 30 untrans-lated region (UTR) of the cDNA [35] We identified aSECIS element in the 30 UTR of SgD3 located betweennucleotide positions 1267 and 1367, but we failed toidentify a SECIS element in SgD2 and may therefore havesub-cloned a fragment without a 30UTR and poly (A) sig-nal in this study The difficulty in obtaining such a frag-ment may be partially attributable to the occurrence of longintrons; it has been reported in mammals that the 30UTR ofD2 has long introns (8.1–8.5 kb) within two exons [36].Similar incomplete sequences have been reported for D2 ofthe Senegalese sole [37] and mummichog [38] and con-sidered to be due to either the lack of an extended 30UTR(up to 7.5 kb in length) including the SECIS structure [39]

or the fragment being a splice variant [37] The fullsequence of mummichog D2 cDNA was later cloned and aSECIS element was found within the 4652 bp region with

an intron divided by a 4.8 kb exon [40]

SgD2 mRNA was highly expressed in the liver, skin,brain, and spleen, and SgD3 mRNA was highly expressed

in the liver, retina, brain, and skin, although the expression

of both genes was, to some extent, detected in all of thetissues tested These results suggest that the organs and

amino acid sequence of the S.

guttatus iodothyronine

deiodinase type III (SgD3)

sequence of the full-length

cDNA clone The complete

mRNA spans 1492 bp with an

ORF of 804 bp (269 amino

acids) Sec in bold

Selenocysteine residue at

position 504 The start and stop

codons are indicated in bold,

with the stop codon denoted by

an asterisk The putative SECIS

element determined by the

tissues with a high expression of SgD2 and SgD3 play a

role in metabolism of thyroid hormones Exceptionally

high SgD3 mRNA levels were detected in the retina,

fol-lowed by the brain The transcription pattern of both genes

was different in the respective tissues, as has also been

observed in some mammals and birds [41,48] and fishes

[20,22] where the deiodinases were found to be regulated

in relation to growth and development, hormonal

treat-ment, thyroid status, pollution biomarkers, and food

availability A simultaneous comparison of D2 and D3

mRNA levels has been carried out in walleye using reverse

transcription-PCR; D2 mRNA abundance in the liver was

significantly higher than in all other tissues, while D3

mRNA was highly expressed in the liver and whole eye,

followed by the brain, gills, and skin [20] In terms of

deiodination activities, high levels of T4ORD and T4IRD

were observed in the liver and brain, respectively, of the

blue tilapia O aureus [42], salmonids [43], and Atlantic

cod [14] In rainbow trout under physiological conditions,

the predominant deiodinase pathways in the brain were

observed to autoregulate T3levels through the degradation

of T4and T3, while the liver generated T3[12] A positive

correlation between hepatic D2 activity and plasma T3

levels has been found in the Nile tilapia [27] and red drum

Sciaenops ocellatus [44] This situation appears to varyamong vertebrates; D1 inactivation and D3 activationcoincidentally occur in the mammalian liver [45–47], whilethe activation of hepatic D3 is one of the main factorsresponsible for decreasing plasma T3 levels in chicken[48]

To the best of our knowledge, daily variations of thyronine deiodinase transcript levels have only beenreported in the late metamorphic stages of the Senegalesesole [37]; D3 transcript levels in larval homogenates weremeasured using qPCR and found to significantly increasefrom Zeitgeber time (ZT) 7 to ZT12 and then decreasefrom ZT12 to ZT24 The results of our study clearly showthat the abundance of SgD2 and SgD3 mRNA in severaltissues was higher at 1200 hours than at 2400 hours, sug-gesting that TH levels fluctuate daily In contrast to T4,little or no daily variation in plasma T3 levels has beenreported in certain teleosts [49,50] However, the plasmalevels of total triiodothyronine (TT3) were found toincrease during the scotophase in juvenile Atlantic salmonparr when they were reared under LD = 8:16 in the winter,while in the spring the TT3 levels were higher in smolts,but there was no daily rhythm The opposite effect wasobserved for total thyroxine in parrs and smolts [51] Daily

iodothyronine deiodinase gene

abundance a SgD2 and b SgD3

mRNA expression in goldlined

spinefoot kept under natural

conditions for 1 week and

sampled at 1200 hours (white

bars) and at 2400 hours (black

bars) Data are given as the

mean ± standard error of the

mean (SEM) (n = 7/group).

Asterisks significant differences

according to Student’s t test

(P \ 0.05)

fluctuations of plasma T4 and T3 levels has also been

reported in juvenile red drum, with an increase during the

photophase in fish fed 1 h before the light were turned off,

dusk-fed fish, and dawn-fed fish kept under LD = 12:12 at

23°C [42] Similar increases in plasma TH levels during

the photophase were observed in goldfish Carassius

auratus reared under LD = 12:12 [52] and in channel

catfish Ictalurus punctatus reared under a natural

photo-period in July [53] A free-running circadian rhythm of

circulating T4 levels was also noted in the juvenile red

drum reared under constant photoperiod conditions with

and without feeding [54] These findings imply that an

endogenous circadian clock regulates TH levels

Concur-rent variations of T3with T4in certain teleost species may

mean that the activity of D2 and D3 is regulated by the

circadian system and that it influences the intercellular and

extracellular levels of TH Our data may imply that

mel-atonin is directly or indirectly related to daily variations in

SgD2 and SgD3 mRNA in the liver because this hormone

increased during nighttime (peak at 2400 hours) anddecreased during daytime In fact, the expression of SgD2

in the hypothalamus was down-regulated by melatoninadministration (Wambiji, Hur, Takeuchi, and Takemuraunpublished data) Similar effects of melatonin on theexpression of iodothyronine deiodinase genes may occur inthe liver

Our results demonstrate that food deprivation loweredplasma glucose and hepatic glycogen levels, similar toresults reported for channel catfish [55], suggesting thatstored nutrients in the liver are mobilized after fooddeprivation Concomitant with these metabolic changes,the mRNA abundance of hepatic SgD2 decreased follow-ing food deprivation It was reported that hepatic D2activity decreases after starvation, with concurrent increa-ses in plasma T3 levels and hepatic D2 activity after re-feeding [27] Food deprivation was observed to lower the

spinefoot Hepatosomatic index (a) and plasma glucose levels (b) in

fed and unfed fish (n = 7 per group) after a 1-week experimental

period The liver was collected from the fish at 1200 hours Data are

given as the mean ± SEM Asterisks significant differences according

to Student’s t test (P \ 0.05)

abundance in the liver of goldlined spinefoot SgD2 (a) and SgD3 (b) mRNA expression levels in fed and unfed fish after a 1-week experimental period The liver was collected from the fish at

1200 hours and measured for SgD2 and SgD3 mRNA abundance by qPCR Data are given as the mean ± SEM (n = 7/group) Asterisks significant differences according to Student’s t test (P \ 0.05)

plasma TH levels of Nile tilapia [27, 56], rainbow trout

[57], and red drum [42], suggesting that nutrition impacts

on gene transcription via the intracellular and extracellular

levels of TH SgD3 mRNA abundance did not change after

food deprivation This results seem to be compatible with

the case of starved tilapia in which D3 activity also

decreased in the gill, brain, and liver [27]

In our study, SgD2 and SgD3 mRNA levels were

affected by temperature, with the highest expression levels

of both genes being observed in fish reared under an

intermediate temperature (25°C), but not at lower (20°C) or

higher (30°C) temperatures Since our study parameters

mimic the minimum winter and maximum summer

tem-peratures, the abundance of these genes in the liver is likely

to reflect their physiological responses In this regard, there

was a crucial effect of temperature on the growth of the

juvenile marbled spinefoot (S rivulatus), a relative species

to the goldlined spinefoot; when the fish were reared at 17,

22, 27, and 32°C, fish reared at 27 and 32°C were icantly larger In addition, the specific growth rate washigher in fish reared at 27 than at 32°C [58] A similartemperature effect was observed in the reproductiveactivity of the sapphire devil Chrysiptera cyanea, a com-mon species in the coral reefs of the West Pacific Ocean[59,60] Overall, these results show that tropical fish have

signif-a suitsignif-able rsignif-ange of tempersignif-ature for signif-an optimsignif-al cal state, including the deiodination activities in the liver.Based on these results, we conclude that the abundance

physiologi-of SgD2 and SgD3 mRNA in the spinefoot is subject to theexogenous factors they are exposed to It is possible thatthe impacts of exogenous factors are transduced to the liverthrough endogenous factors, such as melatonin for day–night difference and growth hormone–insulin-like growthfactor for nutrition status Further studies are needed toelucidate the involvement of endogenous factors in thealternation of SgD2 and SgD3 mRNA levels in the gold-lined spinefoot

Grant-in-Aid for Scientific Research from the Japan Society for the Promotion

of Science (JSPS) and a Joint Research Project under the Japan–Korea Basic Scientific Cooperation Program from JSPS to AT.

4 St Germain DL, Galton VA (1997) The deiodinase family of selenoproteins Thyroid 7:655–668

5 Mol KA, Van der Geyten S, Burel C, Ku¨hn ER, Boujard T, Darras VM (1998) Comparative study of iodothyronine outer ring and inner ring deiodinase activities in five teleostean fishes Fish Physiol Biochem 18:253–266

6 Garcı´a-G C, Jeziorski MC, Valverde-R C, Orozco A (2004) Effects of iodothyronines on the hepatic outer-ring deiodinating pathway in killifish Gen Comp Endocrinol 135:201–209

7 Bianco AC, Salvatore D, Gereben B, Berry MJ, Larsen PR (2002) Biochemistry, cellular and molecular biology, and physiological roles of the iodothyronine selenodeiodinases Endocr Rev 23:38–89

8 Salvatore D, Low SC, Berry MJ, Maia AL, Harney JW, Croteau

W, St Germain D, Larsen PR (1995) Type 3 iodothyronine iodinase: cloning, in vitro expression, and functional analysis of the placental selenoenzyme J Clin Invest 96:2421–2430

de-9 Van der Geyten S, Van den Eynde I, Segers IB, Ku¨hn ER, Darras

VM (2002) Differential expression of iodothyronine deiodinases

in chicken tissues during the last week of embryonic ment Gen Comp Endocrinol 128:65–73

deiodinase gene abundance in the liver of goldlined spinefoot.

SgD2 (a) and SgD3 (b) mRNA expression levels in fish kept at 20, 25,

and 30°C under LD = 12:12 for 1 week The liver was collected from

the fish at 1200 hours and subjected measured for SgD2 and SgD3

mRNA abundance by qPCR Data are given as the mean ± SEM.

n = 7/group Different letters indicate significant differences after

one-way ANOVA (P \ 0.01)

10 MacLatchy DL, Eales JG (1992) Properties of T450-deiodinating

systems in various tissues of the rainbow trout, Oncorhynchus

mykiss Gen Comp Endocrinol 86:313–322

11 Mol K, Kaptein E, Darras VM, de Greef WJ, Ku¨hn ER, Visser TJ

(1993) Different thyroid hormone-deiodinating enzymes in tilapia

(Oreochromis niloticus) liver and kidney FEBS Lett 321:140–144

12 Frith SD, Eales JG (1996) Thyroid hormone deiodination

path-ways in brain and liver of rainbow trout, Oncorhynchus mykiss.

Gen Comp Endocrinol 101:323–332

13 Moore VanPutte CL, MacKenzie DS, Eales JG (2001)

(Sciaenops ocellatus) Comp Biochem Physiol B 128:413–423

14 Cyr DG, Idler DR, Audet C, McLeese JM, Eales JG (1998)

Effects of long-term temperature acclimation on thyroid hormone

deiodinase function, plasma thyroid hormone levels, growth, and

reproductive status of male Atlantic cod, Gadus morhua Gen

Comp Endocrinol 109:24–36

15 Orozco A, Linser P, Valverde-R C (2000) Kinetic

characteriza-tion of outer-ring deiodinase activity (ORD) in the liver, gill and

retina of the killifish Fundulus heteroclitus Comp Biochem

Physiol B 126:283–290

16 Morin PP, Hara TJ, Eales JG (1993) Thyroid hormone

deiodin-ation in brain, liver, gill, heart and muscle of Atlantic salmon

transformation I Outer- and inner-ring thyroxine deiodination.

Gen Comp Endocrinol 90:142–156

17 Johnston CE, Eales JG (1995) Effects of acclimation and assay

trout, Oncorhynchus mykiss J Exp Zool 272:426–434

18 Specker JL, Eales JG, Tagawa M, Tyler WA III (2000)

Parr-smolt transformation in Atlantic salmon: thyroid hormone

deio-dination in liver and brain and endocrine correlates of change in

rheotactic behavior Can J Zool 78:696–705

19 Plohman JC, Dick TA, Eales JG (2002) Thyroid of lake sturgeon,

Acipenser fulvescens II Deiodination properties, distribution, and

effects of diet, growth, and a T3 challenge Gen Comp Endocrinol

125:56–66

20 Picard-Aitken M, Fournier H, Pariseau R, Marcogliese DJ, Cyr

DG (2007) Thyroid disruption in walleye (Sander vitreus)

exposed to environmental contaminants: cloning and use of

iodothyronine deiodinases as molecular biomarkers Aquat

Tox-icol 83:200–211

21 Adams BA, Cyr DG, Eales JG (2000) Thyroid hormone

deio-dination in tissues of American plaice, Hippoglossoides

polychlorinated biphenyls (PCBs) 77 and 126 Comp Biochem

Physiol C 127:367–378

22 Sanders JP, Van der Geyten S, Kaptein E, Darras VM, Ku¨hn ER,

Leonard JL, Visser TJ (1999) Cloning and characterization of

type III iodothyronine deiodinase from the fish Oreochromis

niloticus Endocrinology 140:3666–3673

ration on hepatic deiodination and conjugation of thyroid

hor-mones in rainbow trout Oncorhynchus mykiss Gen Comp

Endocrinol 115:379–386

24 Mol KA, Van der Geyten S, Ku¨hn ER, Darras VM (1999) Effects

of experimental hypo- and hyperthyroidism on iodothyronine

deiodinases in Nile tilapia, Oreochromis niloticus Fish Physiol

Biochem 20:201–207

25 Van der Geyten S, Byamungu N, Reyns GE, Ku¨hn ER, Darras

VM (2005) Iodothyronine deiodinases and the control of plasma

and tissue thyroid hormone levels in hyperthyroid tilapia

(Ore-ochromis niloticus) J Endocrinol 184:467–479

26 Plate EM, Adams BA, Allison WT, Martens G, Hawryshyn CW,

Eales JG (2002) The effects on thyroxine or a GnRH analogue on

thyroid hormone deiodination in the olfactory epithelium and retina

of rainbow trout, Oncorhynchus mykiss, and sockeye salmon, corhynchus mykiss nerka Gen Comp Endocrinol 127:59–65

On-27 Van der Geyten S, Mol KA, Pluymers W, Ku¨hn ER, Darras VM

(Oreochromis niloticus) are mainly regulated through changes in hepatic type II iodothyronine deiodinase Fish Physiol Biochem 19:135–143

28 Walpita CN, Grommen SVH, Darras VM, Van der Geyten S (2007) The influence of stress on thyroid hormone production and peripheral deiodination in the Nile tilapia Oreochromis niloticus Gen Comp Endocrinol 150:18–25

29 Cyr DG, Eales JG (1988) In vitro effects of thyroid hormones on gonadotropin-induced estradiol-17b secretion by ovarian follicles

of rainbow trout, Salmo gairdneri Gen Comp Endocrinol 69:80–87

30 MacLatchy DL, Eales JG (1988) Short-term treatment with

fontinalis Gen Comp Endocrinol 71:10–16

31 MacLatchy DL, Eales JG (1990) Growth hormone stimulates

-triio-dothyronine levels in rainbow trout (Salmo gairdneri) Gen Comp Endocrinol 78:164–172

32 Sambroni E, Gutieres S, Cauty C, Guiguen Y, Breton B, Lareyre

JJ (2001) Type II iodothyronine deiodinase is preferentially expressed in rainbow trout (Oncorhynchus mykiss) liver and gonads Mol Reprod Dev 60:338–350

33 Park YJ, Park JG, Hiyakawa N, Lee YD, Kim SJ, Takemura A (2007) Diurnal and circadian regulation of a melatonin receptor, MT1, in the golden rabbitfish, Siganus guttatus Gen Comp Endocrinol 150:253–262

34 Livak KJ, Schmittgen TD (2001) Analysis of relative gene

37 Isorna E, Obregon MJ, Calvo RM, Va´zquez R, Pendo´n C, Falco´n

J, Mun˜oz-Cueto JA (2009) Iodothyronine deiodinases and thyroid hormone receptors regulation during flatfish (Solea senegalensis) metamorphosis J Exp Zool (Mol Dev Evol) 312B:231–246

38 Valverde-R C, Croteau W, LaFleur Jr GJ, Orozco A, St Germain

deio-dinase from the liver of Fundulus heteroclitus Endocrinology 138:642–648

39 Davey JC, Schneider MJ, Becker KB, Galton VA (1999) Cloning

of a 5.8 kb cDNA for a type 2 deiodinase Endocrinology 140:1022–1025

40 Orozco A, Jeziorski MC, Linser PJ, Greenberg RM, Valverde-R

C (2002) Cloning of the gene and complete cDNA encoding a type 2 deiodinase in Fundulus heteroclitus Gen Comp Endocri- nol 128:162–167

41 Wagner MS, Morimoto R, Dora JM, Benneman A, Pavan R, Maia

AL (2003) Hypothyroidism induces type 2 iodothyronine dinase expression in mouse heart and testis J Mol Endocrinol 31:541–550

deio-42 Mol KA, Van der Geyten S, Darras VM, Visser TJ, Ku¨hn ER (1997) Characterization of iodothyronine outer ring and inner ring deiodinase activities in the blue tilapia, Oreochromis aureus Endocrinology 138:1787–1793

43 Eales JG, MacLatchy DL, Sweeting RM (1993) Thyroid hormone deiodinase systems in salmonids, and their involvement in the regulation of thyroidal status Fish Physiol Biochem 11:313–321

44 Leiner KA, Han GS, MacKenzie DS (2000) The effects of

pho-toperiod and feeding on the diurnal rhythm of circulating thyroid

hormones in the red drum, Sciaenops ocellatus Gen Comp

Endocrinol 120:88–98

45 Chopra IJ (1980) Alterations in monodeiodination of

iodothyro-nines in the fasting rat: effects of reduced nonprotein sulfhydryl

groups and hypothyroidism Metabolism 29:161–167

46 Chopra IJ, Wiersinga W, Frank H (1981) Alterations in hepatic

monodeiodination of iodothyronines in the diabetic rat Life Sci

28:1765–1776

47 O’Mara BA, Dittrich W, Lauterio TJ, St Germain DL (1993)

hormones and in fasted and diabetic rats Endocrinology

133:1715–1723

48 Darras VM, Mol KA, Van der Geyten S, Ku¨hn ER (1998) Control

of peripheral thyroid hormone levels by activating and

inacti-vating deiodinases Ann NY Acad Sci 839:80–86

49 Gomez JM, Boujard T, Boeuf G, Solari A, Le Bail PY (1997)

Individual diurnal plasma profiles of thyroid hormones in

rain-bow trout (Oncorhynchus mykiss) in relation to cortisol, growth

hormone, and growth rate Gen Comp Endocrinol 107:74–83

50 Leiner KA, MacKenzie DS (2003) Central regulation of thyroidal

51 Ebbesson LOE, Bjo¨rnsson BT, Ekstro¨m P, Stefansson SO (2008)

Daily endocrine profiles in parr and smolt Atlantic salmon Comp

Biochem Physiol A 151:698–704

52 Spieler RE, Noeske TA (1981) Timing of a single daily meal and

diel variations of serum thyroxine, triiodothyronine and cortisol

in goldfish Carassius auratus Life Sci 28:2939–2944

53 Loter TC, MacKenzie DS, McLeese J, Eales JG (2007) Seasonal changes in channel catfish thyroid hormones reflect increased magnitude of daily thyroid hormone cycles Aquaculture 262:451–460

54 Leiner KA, MacKenzie DS (2001) The effects of photoperiod on growth rate and circulating thyroid hormone levels in the red drum, Sciaenops ocellatus: evidence for a free-running circadian

55 Shoemaker CA, Klesius PH, Lim C, Yildirim M (2003) Feed deprivation of channel catfish, Ictalurus punctatus (Rafinesque), influences organosomatic indices, chemical composition and sus- ceptibility to Flavobacterium columnare J Fish Dis 26:553–561

56 Toguyeni A, Baroiller JF, Fostier A, Le Bail PY, Ku¨hn ER, Mol

KA, Fauconneau B (1996) Consequences of food restriction on short-term growth variation and on plasma circulating hormones

in Oreochromis niloticus in relation to sex Gen Comp nol 103:167–175

Endocri-57 Flood CG, Eales JG (1983) Effects of starvation and refeeding on

Salmo gairdneri Can J Zool 61:1949–1953

58 Saoud IP, Mohanna C, Ghanawi J (2008) Effects of temperature

on survival and growth of juvenile spinefoot rabbitfish (Siganus rivulatus) Aquac Res 39:491–497

59 Bapary MAJ, Fainuulelei P, Takemura A (2009) Environmental control of gonadal development in the tropical damselfish Chrysiptera cyanea Mar Biol Res 5:462–469

60 Bapary MAJ, Takemura A (2010) Effect of temperature and photoperiod on the reproductive condition and performance of a tropical damselfish Chrysiptera cyanea during different phases of reproductive season Fish Sci 76:769–776

O R I G I N A L A R T I C L E Biology

Migration history of Sakhalin taimen Hucho perryi captured

in the Sea of Okhotsk, northern Japan, using otolith Sr:Ca ratios

Kyoko Suzuki•Tomoyasu Yoshitomi•

Yoichi Kawaguchi•Masaki Ichimura•

Kaneaki Edo• Tsuguo Otake

Received: 24 August 2010 / Accepted: 4 February 2011 / Published online: 16 March 2011

Ó The Japanese Society of Fisheries Science 2011

Abstract Sakhalin taimen Hucho perryi populations have

decreased drastically to near extinction It is urgent to

establish an effective conservation strategy based on an

understanding of the characteristics of migration and

hab-itat use of this species We examined the migration history

of anadromous Sakhalin taimen captured off the Sarufutsu

coast, northern Hokkaido, Japan, using otolith Sr:Ca ratios

and also examined the relationship between their otolith

Sr:Ca ratios during freshwater and seawater residence in a

rearing experiment Otolith Sr:Ca ratios of some fish from

the Sarufutsu coast showed freshwater levels (0.5–4.0 9

10-3) near the core, which thereafter increased to brackish

water levels (4.0–6.0 9 10-3), and then to seawater levels(6.0–10.0 9 10-3) in the outermost regions Those findingsindicate that specimens from the Sarufutsu coast migrated tothe brackish water region or the sea and spent most of theirlives there The anadromous migration pattern including thetiming of downstream migration seems to be flexible amongindividuals in the species They migrate between freshwaterand seawater or brackish water several times during their lives,showing extensive habitat use It is essential to secure thecontinuity among the freshwater, brackish water, and seawa-ter areas for their effective conservation

Keywords Sakhalin taimen
Migration
Rare species
Anadromous
Otolith Sr:Ca ratios
Conservation

IntroductionSakhalin taimen Hucho perryi is a species of the genusHucho, which is composed of five species [1, 2] Thedistribution is limited to the far northeastern part of Asia,from the Primorye region of Siberia to Sakhalin Island, thesouthern Kurile Islands, and the northern area of Hokkaido[1] Unlike the other species of the genus, which are strictlyfreshwater residents, the Sakhalin taimen has been known

to perform an anadromous migration [2 4] The population

of this species has decreased remarkably and is close to theextinction level [5] This situation is thought to result fromvarious human impacts such as indiscriminate fishing,water pollution by development of agricultural land, andhabitat destruction Takami et al [6] reported that thedecrease of Sakhalin taimen was probably the result of theloss of riparian forests and riverine habitats associated withextensive development of agricultural land during the1960s and 1970s The construction of dams, barrages, and

The author Kyoko Suzuki is Research Fellow of the Japan Society for

the Promotion of Science.

International Coastal Research Center,

Atmosphere and Ocean Research Institute,

The University of Tokyo, 2-106-1 Akahama,

Otsuchi, Iwate 028-1102, Japan

e-mail: k-suzuki@aori.u-tokyo.ac.jp

T Yoshitomi

Field Studies Institute for Environmental Education,

Tokyo Gakugei University, 4-1-1 Nukuikitamachi,

Koganei, Tokyo 184-8501, Japan

Y Kawaguchi

Division of Ecosystem Design, Institute of Technology

and Science, The University of Tokushima,

2-1 Minami-josanjima, Tokushima 880-8506, Japan

M Ichimura

Shibetsu Salmon Museum, 1-1 Kita 1-Jo Nishi 6-Chome,

Shibetsu, Shibetsu-gun, Hokkaido 086-1631, Japan

K Edo

Monuments and Sites Division, Agency for Cultural Affairs,

Marunouchi 2-5-1, Chiyoda-ku, Tokyo 100-8959, Japan

DOI 10.1007/s12562-011-0335-x

banks also caused habitat destruction including the loss of

spawning grounds In particular, the straightening of the

rivers at middle to lower reaches conducted in the last

several decades has made the rivers markedly monotonous

with a loss of rapids and deep pools, and has damaged the

river environment for the reproduction of H perryi [7] The

life history, including migration and habitat use of this

species, should be urgently studied for the establishment of

an effective conservation strategy

However, information about the life history of

anadro-mous Sakhalin taimen is still limited with only a few

studies being done on ecological aspects of their freshwater

phases [8 11], the morphology of anadromous Sakhalin

taimen [12], or on the microchemistry of the otoliths and

scales [13–15] Fish otoliths are metabolically inert with

the aragonite mineralogy remaining unaltered after

depo-sition [16], so the elemental compodepo-sition of the otolith

reflects to some degree the environment of the water in

which the fish lives [17] The strontium content in otoliths,

in particular, varies with fluctuations in ambient salinity,

allowing the reconstruction of the anadromous migration

history of each fish [18–24] Regarding Sakhalin taimen,

Arai et al [13] used otolith analysis to report downstream

migration of Sakhalin taimen collected at Lake Aynskoye

in Sakhalin Island Honda et al [14] analyzed otolith Sr:Ca

ratios of Sakhalin taimen caught from Lake Akkeshi

(brackish water lake) in Hokkaido, Japan, and suggested

that the specimens had migrated into brackish waters, but it

was unlikely they went into the ocean The present study is

the first report on detailed migration history of anadromous

Sakhalin taimen, actually captured in the Sea of Okhotsk,

the Sarufutsu coast, northern Hokkaido in Japan The

population of Sakhalin taimen in the Sarufutsu River

sys-tem is one of the stable populations of this species in Japan

[25] Furthermore, there seems to be a genetic difference

among the stocks of several river systems (K Edo, pers

comm., 2007, 2009) Therefore the life history and migration

traits of Sakhalin taimen should be investigated in each river

system unit This knowledge is vital for establishment of an

effective conservation strategy for the species whose local

population is declining The Sr:Ca ratio levels of Sakhalin

taimen otoliths from fish reared in seawater and freshwater

were also examined to verify the affect of salinity on the

Sr:Ca ratios in the otoliths of this species

Materials and methods

Rearing experiment

A total of 12 Sakhalin taimen used for this experiment were

provided by the Shibetsu Salmon Museum We do not

report the river lineage from which these specimens were

taken to help protect this critically endangered species.They were artificially hatched and kept in freshwater at16.0–17.0°C under natural light conditions for 6 monthsbefore the experiment The average fork length and bodyweight at the start of the rearing experiment were

424 ± 41 mm and 838 ± 250 g, respectively The rearingexperiment was carried out in the Shibetsu SalmonMuseum Six of 12 fish remained in a freshwater envi-ronment for another 6 months, and the other six weretransferred to seawater and also reared for 6 months Thefreshwater fish were reared in a 500 L tank supplied withwell water at 10–20 L/min The temperature of the wellwater was ca 16.8°C The seawater rearing was in filtrationtank (5,000 L) of the Shibetsu Salmon Museum, which has

a water recycling system The seawater used for the rearingwas pumped up from the Shibetsu port and circulated in thesystem at the rate of 180–240 L/min The salinity of sea-water used for the rearing was 30 psu The seawater tem-perature varied from 12 to 17°C, and the averagetemperature was ca 15°C during the experiment Lightingconditions of each rearing tank were based on natural light.Fish of both freshwater and seawater groups were fed

4 days per week with dry pellets (Hokuren) We could notcarry out the rearing experiment under a variety of salini-ties because of the limited numbers of this locally protectedfish species

Wild fish used for the analyses of otolith Sr:Ca ratio

A total of seven wild Sakhalin taimen were used for theanalysis of otolith Sr:Ca ratios All fish were unintention-ally caught in a set net fishery mainly targeting pink salmon(Oncorhynchus gorbuscha), chum salmon (Oncorhynchusketa), and masu salmon (Oncorhynchus masou) conducted

by the Sarufutsu Fishery Cooperative and Fujimoto eries Company along the Sarufutsu coast of northernHokkaido, Japan (the Sea of Okhotsk) (Fig.1)

Fish-The fork length, body weight, sex, age, capture date, andnumber of retained ovulated eggs of each fish are shown inTable1 Scales were used for the age determination ofeach fish Scales were removed from areas above andbelow the lateral line of the fish body

Otolith preparation and Sr:Ca ratio analysesSagittal otoliths were extracted from each fish and wereembedded in epoxy resin (Epofix, Struers, Denmark) Theywere mounted on glass slides by epoxy bond and ground toexpose the core using a grinding machine equipped with adiamond-cup wheel of 13 lm (Discoplan-TS, Struers,Denmark) The ground surface of otolith was polished with

6 and 1 lm diamond paste on a polishing wheel

(Planopol-V equipped with PdM-Force, Struers, Denmark), Pt-Pd

coated with a high vacuum evaporator (E-1030, Hitachi,

Japan) after washing with deionized water

For life history transect analysis, the profiles of Sr and

Ca concentrations were analyzed from the core to the edge

along the radius of each otolith using a wave-length

dis-persive X-ray electron probe microanalyzer (EPMA; JEOL

JXA-8900, Jeol, Japan) The accelerating voltage and beam

current were 15 kV and 1.2 9 10-8A, respectively The

electron beam for the rearing experiment samples was

focused on a point of 10 lm diameter, and the

measure-ments were spaced at 10 lm intervals The beam for wild

fish was focused on a point of 5 lm diameter, and the

measurements were spaced at 5 lm intervals The

outer-most 100 lm of the otolith of fish no 3 (Table1) was

further examined to determine the Sr:Ca ratio deposited

under seawater conditions For this analysis beam diameter

and measurement interval were both set to 1 lm with the

accelerating voltage and beam current the same as

men-tioned above

Two-dimensional X-ray intensity maps of Sr and Ca

were also examined by EPMA under the following

measurement conditions: accelerating voltage, 15 kV;beam current, 5.0 9 10-7A; pixel size, 10 9 10 lm.Calcite (CaCO3) and strontianite (SrCO3) were used asstandards

ResultsThe rearing experiment showed that there was a remark-able difference in the Sr:Ca ratios between the otoliths ofthe fish reared in the freshwater and seawater tanks(Fig.2) The otolith Sr:Ca ratios of individuals reared infreshwater (freshwater sample) maintained lower levelswith an average ratio of 1.1 9 10-3(range 0–3.4 9 10-3)throughout the rearing period In contrast, the ratios ofindividuals transferred into seawater from freshwater(seawater sample) sharply increased to a high level with anaverage of 5.6 9 10-3 (range 4.0–7.8 9 10-3) at about

500 lm from the otolith edge These facts are clearlysupported by two-dimensional X-ray intensity maps of Srcontent for otoliths of freshwater and seawater samples

individual Sakhalin taimen

captured on the coast of

Sarufutsu

(years)

Fork length (cm)

Body weight (g)

Number of retained eggs

(Fig.3) These results indicated that otolith Sr:Ca ratios of

less than 3.0 9 10-3 and more than 5.0 9 10-3 can be

correlated with the freshwater and seawater living phases

of individual fish, respectively, with ratios between those

two levels being associated with the estuarine living phase

Sr:Ca ratio profiles along the otoliths of each fish from the

Sarufutsu coast (fish nos 1–7) are shown in Fig.4 There

appeared to be two or three phases in most of the profiles,

with the inner regions closer to the core being lower thanthe outer levels Some otoliths went from lower values tohigher values (nos 1, 2, 3, 4, 5) Others went from higher

to lower values (nos 6 and 7) One, no 6, initially dropped,remained relatively low, and then shifted higher The innerregion with the lower ratio, which extended 1,000 lm fromthe core in fish nos 1–4 and 4,000 lm in fish nos 5 and 7,possibly corresponded to the freshwater living period of the

in the outer edge of the otolith

(1,000 lm from the edge) of

fish reared in freshwater (a,

b) and seawater (c, d) The

otolith Sr:Ca ratios of

individuals reared in freshwater

remained low at the edge of

otolith In contrast, the ratios of

individuals transferred into

seawater from freshwater

sharply increased to a high level

at about 500 lm from the

otolith edge

the Sr contents in the otoliths of

fish reared in freshwater and

seawater Freshwater sample

showed low Sr levels (blue)

consistently Seawater sample

showed low Sr levels (blue) at

the inside and high Sr levels

(green) at the outside

alevin stage before the downstream migration The Sr:Ca

ratio in the outermost portion of the no 3 fish otolith

(20 lm from the edge) corresponded to the seawater living

period and averaged 7.2 9 10-3(range 5.6–10.1 9 10-3),

which suggests that the otolith Sr:Ca ratio under natural

conditions agreed with the ratio under rearing conditions

(Fig.5)

Discussion

We examined the migration history of anadromous

Sakha-lin taimen captured along the Sarufutsu coast, northern

Hokkaido, Japan, using otolith Sr:Ca ratios We also

examined the relationship between the otolith Sr:Ca ratios

of captive fish reared in freshwater and seawater conditions

The otolith Sr:Ca ratios that corresponded to the rearing

period in freshwater and seawater were 1.1 9 10-3(range

0–3.4 9 10-3) and 5.6 9 10-3 (range 4.0–7.8 9 10-3),

respectively The otolith Sr:Ca ratios of brackish water wereestimated to be 3.0–5.0 9 10-3, which agreed with therange reported by Honda et al [14] and Arai [26] In thesestudies, otolith Sr:Ca ratios of some wild fish from the Sa-rufutsu coast showed low levels (0.5–4.0 9 10-3) near thecore, which thereafter greatly increased to 4.0–6.0 9 10-3with the ratios being at higher levels (6.0–10.0 9 10-3) inthe outermost regions Those three levels correspond to thefreshwater, brackish water, and seawater living periods,respectively These findings suggest that specimens fromthe Sarufutsu coast migrated to the brackish water region orthe sea of Okhotsk and spent most of their lives there Onthe other hand, Honda et al [14] analyzed the otolith Sr:Caratios of Sakhalin taimen caught from Lake Akkeshi(brackish water lake) in Hokkaido, Japan, and suggested thespecimens had migrated into brackish waters, but it wasunlikely that they went into the ocean The variety in the use

of the ocean and brackish water habitat between fish fromthe Sarufutsu coast and Lake Akkeshi seems to be related tothe geographical differences in those estuaries Estuariesalong the Sarufutsu coast open directly to the ocean, whileLake Akkeshi is connected to the ocean through AkkeshiBay Furthermore, a genetic difference has been suggestedbetween the Sakhalin taimen of the Sarufutsu River systemand the Bekanbeushi River system, which flows into LakeAkkeshi (K Edo, pers comm., 2007, 2009) Genetic dif-ferences might affect the migration pattern and life historytraits of those stocks

The large difference in the pattern of Sr:Ca ratios ofotoliths of wild fish suggests that there are two timings inthe initiation of downstream migration in Sakhalin taimen.Fish nos 1–4 seemed to move down to the estuary or seabefore age 3 since their otolith radius of 1,000 lm was

otoliths of wild fish collected from the Sarufutsu Coast

(110 lm from the edge) of fish no 3, which was collected from the Sarufutsu Coast

smaller than those of reared fish aged 2? (K Suzuki,

un-publ data, 2009) On the other hand, fish nos 5 and 7, in

which Sr:Ca ratios increased at an otolith radius of ca

4,000 lm, possibly performed seaward migration after the

age of 4 years This suggests that the onset of downstream

migration in Sakhalin taimen is quite flexible Sakhalin

taimen have been known to begin smoltification in spring

at age 2? and develop a strong salinity tolerance by

Sep-tember of that year (T Kubo and S Yamashiro, pers

comm., 1982), which is directly linked with the initiation

of downstream migration Honda et al [27] suggested that

Sakhalin taimen moved downstream in the river in spring

and autumn, and the timing was related to individual

metabolic efficiency and abundance of food available in the

stream It is possible that the timing of the smoltification

and seaward migration of anadromous Sakhalin taimen in

the Sarufutsu coast may also be related to various

biolog-ical factors, such as metabolic efficiency, body growth,

distribution of available food, and habitat density in the

river Metcalfe et al [28] suggested that juvenile Atlantic

salmon (Salmo salar) with higher metabolic rates had an

advantage in terms of social status The juveniles of higher

social status tended to undergo metamorphosis into the

migratory smolt stage at a younger age [29,30], suggesting

that the metabolic rate in the juvenile stage likely affects

the life history traits of individual fish Titus et al [31] also

reported that the habitat density of brown trout (Salmo

trutta) potentially affected the age at first smoltification

The relatively high Sr:Ca ratios at the center of the core

and the subsequent drop that was seen in some of the fish

may be related to early life history events such as

emer-gence and exogenous feeding as Saito et al [32] suggested

for chum salmon (Onchorhyncus keta) The higher ratio in

the otolith core of fish no 6, which was comparable to the

ratios from the sea living period, might be due to the higher

ratio of yolk material of the maternal fish, which may have

spent more time in the higher salinity environment during

the maturation Kalish [33] reported that otolith Sr:Ca

ratios in the primordia of the progeny of anadromous

sal-monids were greater than those of the progeny of

nona-nadromous individuals

In the outer portion of the otoliths with high Sr:Ca ratios

in fish nos 1–4—the portion that appears to correspond to

the sea living phase—there were several short drops to

ratios less than 5 9 10-3(Fig.4) These drops suggest that

the fish moved into estuarine environments several times

during the sea living period, or even that they might have

entered freshwater Two-dimensional X-ray intensity map

of the Sr content for the otoliths of fish no 3 reveals several

rings indicating a shift in the living environment from

brackish water to seawater (Fig.6) The fish seem to

migrate between freshwater and seawater or brackish water

several times during their lives

One factor that makes it difficult for Sr:Ca ratios to beused to observe short-term movements in mobile fishessuch as Sakhalin taimen is that there is likely to be a lageffect of elemental uptake into otoliths In black bream(Acanthopagrus butcheri), otolith Sr:Ca ratios onlyincreased with increasing ambient Sr concentrations after

20 days of exposure [34] Fish no 2, captured in the coastalarea on 26 June 1997, had two ovulated ova retained in theabdominal cavity, suggesting that this individual hadspawned in the headwaters of a river less than 2 monthsbefore being caught, since the spawning of Sakhalin taimen

in the coast of Hokkaido only occurs from late April toearly May [7,8] The Sr:Ca ratios in the outermost portion

of the otolith of that individual, however, were about

4 9 10-3, comparable to the estuarine level Sakhalin imen are reported to stay in upstream tributaries for only5–8 days [35] and subsequently to run downstream over1–4 days [27] Such a short stay in a freshwater region mayresult in relatively high Sr:Ca ratios in the outermost region

ta-of otolith with little or no evidence ta-of the short duration inthe freshwater environment for spawning being recorded.The present study confirmed that Sakhalin taimen, cap-tured along the Sarufutsu coast in the Sea of Okhotsk innorthern Hokkaido, appeared to utilize estuarine and coastalareas as alternative growing habitats These results suggestthat it is important to preserve the environment of theestuarine and coastal regions for protection of Sakhalin ta-imen Furthermore the anadromous migration patternincluding the timing of downstream migration may beflexible among individuals in Sakhalin taimen The con-nectivity between headwaters of rivers that are used asspawning grounds of Sakhalin taimen and the estuarine andcoastal sea that may be used as important growth habitats is

from Sarufutsu Coast Sr concentration showed low levels (blue) at the core of the otolith and then higher levels towards the outside (green and yellow) Three yellow concentric rings in the outer portion

of the otolith suggest that the individual moved into a higher salinity region in the estuary or sea

vital for the protection of the species Kawaguchi et al [36]

reported that improvements in inadequately constructed

weirs could restore the migration and spawning of Sakhalin

taimen in the Sarukotsu River in Sarufutsu, which is the

area of the present study Considering the establishment of

the conservation strategy of Sakhalin taimen, it is essential

to maintain the connectivity of the river for the conservation

of the migration route as well as the spawning and growth

habitats In the future, detailed trace element analyses of

otoliths together with other approaches such as

bioteleme-try, which can directly monitor the movement and behavior

of individual fish, would offer the key to understanding the

unknown life history and help to establish an effective

conservation strategy for this endangered species

Conser-vation Group for providing helpful information on H perryi We are

grateful to Sarufutsu Fisheries Cooperative Association and Fujimoto

Fisheries Company for supplying specimens Thanks to Dr Shiro

Sagawa, Aqua Restoration Research Center, Public Works Research

Institute, and Mr Masahiro Kusunoki, Hokkaido Government Souya

General Subprefectural Bureau, for providing information about the

habitat river of H perryi We also thank Dr Michael J Miller for

critical comments and helping to improve the manuscript This study

was supported by a Grant-in-Aid for JSPS Fellows (no 21
10170) to

K.S., and by a Grant-in-Aid for scientific research (no 21380122)

from the Ministry of Education, Culture, Sports, Science, and

Tech-nology of Japan to T.O.K.

References

1 Kimura S (1966) On the life history of the salmonid fish, Hucho

perryi (Brevoort), found in Nemuro, Hokkaido (in Japanese with

English abstract) Jpn J Ichthyol 14:17–25

2 Holcik J, Hensel K, Nieslanik J, Skacel L (1988) The Eurasian

huchen, Hucho hucho: largest salmon of the world Dr W Junk,

Dordrecht

3 Gritsenko OF, Malkin EM, Churikov AA (1974) Sakhalinskii

tai-men, Hucho perryi (Brevoort) reki Bogatoi (vostochnoe poberezh’e

Sakhalina) (in Russian) Izvestiva TINRO 93:91–100 [Japanese

translation by Oya Y (1976) Sakhalin taimen Hucho perryi in

Bogatoi River (eastern Sakhalin) Sakana to Ran 143:25–34]

4 Kawamula H, Mabuchi M, Yonekawa T (1983) The Japanese

huchen, Hucho perryi (Brevoort), collected in brackish water

Lake Akkeshi, eastern Hokkaido, Japan (in Japanese with English

abstract) Sci Rep Hokkaido Fish Hatchery 38:47–55

5 The Japanese Environment Agency (2007) Red list for brackish

and freshwater fishes in Japan Japanese Environment Agency,

Tokyo

6 Takami T, Kawamula H (2008) Decline of the Sakhalin taimen

(Hucho perryi) populations in Hokkaido, Japan Wildl Conserv

Jpn 11:1–5

7 Fukushima M (2001) Salmonid habitat–geomorphology

rela-tionships in low-gradient streams Ecology 82:1238–1246

8 Fukushima M (1994) Spawning migration and redd construction

of Sakhalin taimen, Hucho perryi (Salmonidae) on northern

Hokkaido Island, Japan J Fish Biol 44:877–888

9 Edo K, Kawamula H, Higashi S (2000) The structure and

dimensions of redds and egg pockets of the endangered salmonid,

Sakhalin taimen J Fish Biol 56:890–904

10 Esteve M, McLennan D, Kawahara M (2009) Spawning iour of Sakhalin taimen, Parahucho perryi, from northern Hok- kaido, Japan Environ Biol Fish 85:265–273

behav-11 Nomoto K, Omiya H, Sugimoto T, Akiba K, Edo K, Higashi S (2010) Potential negative impacts of introduced rainbow trout on endangered Sakhalin taimen through redd disturbance in an agri- cultural stream, eastern Hokkaido Ecol Freshw Fish 19:116–126

12 Edo K, Kawaguchi Y, Nunokawa M, Kawamula H, Higashi S (2005) Morphology, stomach contents and growth of the endan- gered salmonid, Sakhalin taimen Hucho perryi, captured in the Sea of Okhotsk, northern Japan: evidence of an anadromous form Environ Biol Fish 74:1–7

13 Arai T, Kotake A, Morita K (2004) Evidence of downstream migration of Sakhalin taimen, Hucho perryi, as revealed by Sr:Ca ratios of otolith Ichthyol Res 51:377–380

14 Honda K, Arai T, Takahashi N, Miyashita K (2010) Life history and migration of Sakhalin taimen, Hucho perryi, caught from Lake Akkeshi in eastern Hokkaido, Japan, as revealed by Sr:Ca

10.1007/s10228-010-0174-2

15 Suzuki K, Yoshitomi T, Kawaguchi Y, Edo K, Homma-Takeda S, Ishikawa T, Iso H, Imaseki H (2008) Application of micro-PIXE analysis for a migration history study of Hucho perryi focused on strontium distribution in fish scales Int J PIXE 18:39–45

16 Campana SE, Neilson JD (1985) Microstructure of fish otoliths Can J Fish Aquat Sci 42:1014–1032

17 Thorrold SR, Jones CM, Campana SE (1997) Response of otolith microchemistry to environmental variations experienced by larval and juvenile Atlantic croaker (Micropogonias undulatus) Limnol Oceanogr 42:102–111

18 Secor DH, Henderson-Arzapalo A, Piccoli PM (1995) Can otolith microchemistry chart patterns of migration and habitat utilization

in anadromous fishes? J Exp Mar Biol Ecol 192:15–33

19 Otake T, Uchida K (1998) Application of otolith microchemistry for distinguishing between amphidromous and non-amphidrom- ous stocked ayu, Plecoglossus altivelis Fish Sci 64:517–521

20 Tsukamoto K, Arai T (2001) Facultative catadromy of the eel Anguilla japonica between freshwater and seawater habitats Mar Ecol Prog Ser 220:265–276

21 Arai T (2002) Migration history of fishes: present status and perspectives of the analytical methods (in Japanese with English abstract) Jpn J Ichthyol 49:1–23

22 Arai T (2007) Studies on life history and migration in fish by

73:652–655

23 Zhong L, Guo H, Shen H, Li X, Tang W, Liu J, Jin J, Mi Y (2007) Preliminary results of Sr:Ca ratios of Coilia nasus in otoliths by micro-PIXE Nucl Instrum Methods Phys Res B 260:349–352

24 Umatani Y, Arai T, Maekawa K (2008) Variation in migratory history of dolly varden in a stream with an artificial dam in the Shiretoko Peninsula, Hokkaido, Japan Environ Biol Fish 83:37–44

25 Edo K (2007) Ecology and conservation of Sakhalin taimen (in Japanese) Nat Hokkaido 45:2–10

26 Arai T (2010) Effect of salinity on strontium:calcium ratios in the otoliths of Sakhalin taimen, Hucho perryi Fish Sci 76:451–455

27 Honda K, Noda Y, Tsuda Y, Yasuma H, Miyashita K (2009) Tracing the seasonal migration of adult Sakhalin taimen, Hucho perryi, using acoustic telemetry (in Japanese with English abstract) Jpn J Ecol 59:239–247

28 Metcalfe NB, Taylor AC, Thorpe JE (1995) Metabolic rate, social status and life-history strategies in Atlantic salmon Anim Behav 49:431–436

29 Metcalfe NB, Huntingford FA, Graham WD, Thorpe JE (1989) Early social status and the development of life-history strategies

in Atlantic salmon Proc R Soc Lond B 236:7–19

30 Metcalfe NB, Huntingford FA, Thorpe JE, Adams CE (1990) The

effects of social status on life-history variation in juvenile

sal-mon Can J Zool 68:2630–2636

31 Titus RG, Mosegaard H (1992) Fluctuating recruitment and

variable life history of migratory brown trout Salmo trutta L., in

a small, unstable stream J Fish Biol 41:239–255

32 Saito T, Kaga T, Seki J, Otake T (2007) Otolith microstructure of

chum salmon Oncorhynchus keta: formation of sea entry check

and daily deposition of otolith increments in seawater conditions.

Fish Sci 73:27–37

33 Kalish JM (1990) Use of otolith microchemistry to distinguish the

progeny of sympatric anadromous and non-anadromous

salmo-nids Fish Bull US 88:657–666

34 Elsdon TS, Gillanders BM (2005) Strontium incorporation into calcified structures: separating the effects of ambient water con- centration and exposure time Mar Ecol Prog Ser 285:233–243

35 Fukushima M (2008) Sakhalin taimen (Hucho perryi): challenges

of saving giant freshwater fish species (in Japanese) Jpn J thyol 55:49–53

Ich-36 Kawaguchi Y, Okamoto M, Kasai M, Okamoto T, Osanai K, Iwase H, Edo K (2007) Improvement of a weir for Sakhalin taimen (Hucho perryi) migration in the Sarukotsu River, northern Japan In: Proceedings of Wild Trout IX Symposium, October 9–12, West Yellowstone, MN, pp 138–142

O R I G I N A L A R T I C L E Biology

Identification of individuals born in different spawning seasons

using otolith microstructure to reveal life history of Neosalanx

taihuensis

Lang Wu•Jia Shou Liu•Xing Lu Wang•

Guo Zhang• Zheng You Zhang• Brian R Murphy•

Song Guang Xie

Received: 28 September 2010 / Accepted: 6 February 2011 / Published online: 2 March 2011

Ó The Japanese Society of Fisheries Science 2011

Abstract Otolith microstructure was used to distinguish

specimens of Neosalanx taihuensis born in spring and

autumn Increment width during the early life stage was

significantly narrower for spring-born than autumn-born

juveniles, and the frequency distributions of the width of

the first 5 increments were distinctive and diagnostic

Otolith growth trajectories and frequency distribution of

the first 5 increments of spring-spawning adults displayed

similar patterns to spring-born juveniles Otolith growth

trajectories of autumn-spawning adults were intermediate

between those of spring- and autumn-born juveniles, and

the frequency distribution of the width of the first 5

increments showed two modes, one similar to spring-born

juveniles and the other similar to autumn-born juveniles

Populations of N taihuensis have previously been shown

to be dominated by spring-spawning fish Thus, the

puta-tive life history of N taihuensis can be summarized as: (1)

a small part of the spring-born fish mature and spawn inautumn of the year of their birth, (2) the majority of thespring-born cohort matures and spawns in spring of thenext year, and (3) offspring of the autumn-born fish matureand spawn in autumn of the following year This stock-structure information should be considered in fisheriesmanagement for this species

Keywords Spring-spawning stock Autumn-spawningstock Identification  Early life history  Otolith growthtrajectories Icefish

IntroductionMany fish populations are composed of multiple stocksfrom different spawning grounds and/or spawning seasons[1 3] This multiple-stock composition ensures that thepopulation can exploit broad resources, and enables thepersistence of the population under changing environ-mental conditions [2,3] Identifying the stock composition

is essential for successful management and conservation ofsuch populations [3 5]

Neosalanx taihuensis is a freshwater icefish dae) mainly inhabiting the middle and lower reaches of theYangtze River, including its tributaries and affiliated lakes[6] Because of its high commercial value, this fish hasbeen introduced into lakes and reservoirs in over 20provinces, and has become the most commerciallyexploited icefish in China [7] In the middle and lowerreaches of the Yangtze River, N taihuensis usually has twospawning stocks, one spawning in autumn from late Sep-tember through early November and the other in springfrom early January through late April [8,9] The life span

(Salangi-of N taihuensis is suggested to be 1 year [9] However, the

Key Laboratory of Biodiversity and Conservation of Aquatic

Organisms, Institute of Hydrobiology, Chinese Academy

of Sciences, Wuhan 430072, Hubei, China

Fisheries Bureau of Danjiangkou City,

Danjiangkou 442700, Hubei, China

B R Murphy

The Department of Fisheries and Wildlife Sciences and

Conservation Management Institute, Virginia Polytechnic

Institute and State University, Blacksburg, VA 24061, USA

DOI 10.1007/s12562-011-0333-z

relationship between the two spawning stocks has never

been thoroughly investigated Are the two spawning stocks

recruited from the offspring of the two stocks

indepen-dently or mixed with each other? The answer to this

question is essential for understanding the fundamental

biology of this species, and also for effective management

of fisheries for this species

Fish otoliths can provide a chronological record of early

life history [10, 11] Daily periodicity of increments has

been validated in otoliths for many fish species during early

life stages [11,12], and width of the daily increment

typ-ically reflects daily growth rate [14, 15] Environmental

variables (e.g., water temperature and food availability) are

important factors influencing early growth and,

conse-quently, growth trajectories in otoliths of fish [13,16] In

the waters along the middle and lower reaches of the

Yangtze River, water temperature in the spring spawning

season is lower than in the autumn spawning season of

N taihuensis, which could induce variations in early

growth and development and, consequently, otolith

microstructure between the fish born in the two seasons

Thus, otolith microstructural analysis is a potential way to

discriminate offspring of the two stocks, and thus

deter-mine the origins of the adults In this study, variations in

otolith growth trajectories of early juveniles of N

taihu-ensis born in both seasons were investigated Otolith

growth trajectories during the early life stage of mature

adults collected in the two spawning seasons were then

compared with the trajectories of early juveniles to

deter-mine their origins

Materials and methods

Study area

Danjiangkou Reservoir (32°300–33°N, 110°400–110°420E),

with an area of 840 km2, is located in the upper reach of

the Hanjiang River, a tributary of the middle reach of the

Yangtze River (Fig.1) N taihuensis was introduced into

the reservoir in 1998 and has become a major commercial

fishery resource there Similar to the populations in otherwaters of the middle and lower reaches of the YangtzeRiver, N taihuensis in Danjiangkou Reservoir has aspring-spawning stock and an autumn-spawning stock Thesurface water temperature in the reservoir was recordedevery day at about 0900 hours from 1 January 2008through 1 January 2009 Water temperature in the springspawning season from January through April ranges from5°C to 16°C with mean of 8.8°C, and is lower than thewater temperature in the autumn spawning season (Sep-tember through November, 7–27.5°C with mean of 20°C)(Fig.2)

Fish sampling

N taihuensis were sampled using a lift net (15 m 9 15 m,4-mm mesh size) Spring- and autumn-born juveniles wereidentified from the specimens sampled following the springand autumn spawning seasons, respectively, on 17 May

2008 and 16 January 2009 Fish were measured for dard length (SL, 0.1 mm) and dissected to check thegonadal development stage The spring-spawning adultswere sampled on 15 March 2009, and the autumn-spawn-ing ones were sampled on 27 September and 18 October

stan-2009 Fish were dissected to determine sex and gonadaldevelopment stage Only females were analyzed in thisstudy Spawning females were identified as fish with ripegonads (gonads at stage IV or V) [9] Such spawningfemales were measured for SL (0.1 mm), then frozen andbrought to laboratory for otolith analysis

Otolith analysisFish were thawed in the laboratory Both the right and leftsagittal otoliths were extracted from each fish, and usuallyonly the left one was analyzed The otoliths were mounted

on a glass slide with nail polish, ground with 1000–1500#waterproof sandpaper, and polished with lapping film(4000#) until the core and the increments were visible bylight microscope The increments were measured along thelongest axis extending from the core to the outmost visible

110 o 40’E 111 o 10 ’ E 111 o 40’ E

20km Danjiangkou Reservoir

increment in the anterior area of otolith using an image

analysis system (Ratoc System Engineering, Tokyo, Japan)

with a direct data feed between the light microscope and a

computer [14] (Fig.4) Increments in sagittal otoliths of

Protosalanx hyalocranius (a related Salangid species) have

been validated as being deposited daily with the first

increment deposited on the second day after hatching [17]

The sagittal otolith of N taihuensis is similar to that of

P hyalocranius in shape and microstructure near the core

area [18] We therefore suggest that increments in the

sagittal otoliths of N taihuensis are likely also deposited

daily, and we performed our analyses based on this

assumption

Data analysis

Analysis of variance (ANOVA) was used to compare the

incremental width and the total width of the first 5

incre-ments among spring- and autumn-born juveniles, and

spring- and autumn-spawning adults Multiple-comparison

Tukey honestly significant difference (HSD) was applied

when the difference was significant (P \ 0.05) Data are

presented as mean ± standard deviation

Results

There were two separated modes of SL frequency

distri-bution for fish sampled on 17 May 2008 and 16 January

2009, respectively For the specimens of 17 May 2008, SL

of the fish in the smaller mode ranged from 27.1 to

43.0 mm with mean of 31.5 ± 2.3 mm (n = 78), and that

in the larger mode ranged from 57.7 to 73.0 mm with mean

of 67.6 ± 4.7 mm (n = 9) (Fig.3a) For the specimens of

16 January 2009, SL of the fish in the smaller mode ranged

from 32.3 to 45.7 mm with mean of 37.8 ± 2.9 mm

(n = 69), and that in the larger mode ranged from 55.6 to69.2 mm with mean of 63.6 ± 3.1 mm (n = 113)(Fig.3b) The gonads of all the smaller-sized fish for bothsamples were at the virgin stage (stage I) The larger-sizedfish in May had gonads at stage VI, and those in Januaryhad gonads at stage IV By reference to reported monthlysizes and cohort compositions of N taihuensis in themiddle Yangtze [9], the smaller mode fish in the specimens

of May 17 were identified as spring-born juveniles, andthose of 16 January were identified as autumn-born juve-niles Spring-spawning adults ranged from 57.8 to74.6 mm SL with mean of 68.5 ± 3.7 mm (n = 101), andautumn-spawning adults ranged from 52.4 to 64.2 mm withmean of 57.8 ± 2.4 mm (n = 82) (Fig.3c, d)

All sagittal otoliths of N taihuensis had a single mordium When observed using a microscope with trans-mitted light, there is a prominent dark zone (D-zone)surrounding the core There were no increments in the corearea inside the D-zone, while there were discernibleincrements outside of the D-zone The second D-zone andthe light zone (L-zone) following formed the first incre-ment There were 40–60 increments discernible bymicroscope outside of the core After that, incrementscould usually not be accurately counted (Fig.4)

Date 08-Jan-1 08-Mar-1 08-May-1 08-Jul-1 08-Sep-1 08-Nov-1 09-Jan-1

0 40 80

0 40 80

0 40 80

taihuen-sis sampled in May 2008 (a, n = 88), January 2009 (b, n = 182), March 2009 (c, n = 101), and September and October 2009 (d, n = 82) in Danjiangkou Reservoir

The otolith-increment growth trajectories showed

dif-ferent patterns between spring- and autumn-born juveniles

For spring-born juveniles, increment width increased

gradually from the first increments of 1.6 ± 0.3 lm (range

1.0–2.3 lm) to the 17th increment of 2.6 ± 0.5 lm (range

1.5–4.2 lm), and remained relatively constant to the 40th

increment For autumn-born juveniles, increment width

increased rapidly from the first increment of 3.0 ± 0.3 lm

(range 2.2–4.9 lm) to the 8th increment of 4.5 ± 0.9 lm

(range 2.6–6.5 lm), and then decreased gradually to

2.9 ± 0.4 lm (range 2.0–3.9 lm) for the 40th increment

Incremental widths from the first through the 38th

incre-ments of spring-born juveniles were significantly narrower

than those of autumn-born juveniles (ANOVA, P \ 0.05;

Fig.5) Growth trajectory for the otolith increments of

spring-spawning adults and larger-sized fish sampled on 17

May 2008 and 16 January 2009 showed a pattern similar to

spring-born juveniles, while the otolith growth trajectory of

autumn-spawning adults was intermediate between those of

spring- and autumn-born juveniles (Fig.5)

The frequency distributions of the total width of the first

5 increments did not overlap between spring- and

autumn-born juveniles The total width of the first 5 increments

ranged from 5.6 to 11.4 lm (mean 8.3 ± 1.3 lm; n = 79)

for spring-born juveniles, and from 12.5 to 22.9 lm (mean

17.4 ± 2.1 lm; n = 68) for autumn-born juveniles Thus,

for the specimens in our study, the birth season of a fish

could be positively identified by measuring the total width

of the first 5 increments The frequency distribution of the

total width of the first 5 increments for spring-spawning

adults and larger-sized fish sampled on 17 May 2008 and

16 January 2009 showed a mode similar to that of

spring-born juveniles The total width of the first 5 increments

ranged from 7.0 to 11.8 lm (mean 8.9 ± 1.1 lm; n =101) for spring-spawning adults, and from 8.5 to 9.5 lm(mean 8.8 ± 0.3 lm; n = 9) and 7.2 to 11.9 lm (mean9.3 ± 1.1 lm; n = 87) for larger-sized fish sampled on 17May 2008 and 16 January 2009, respectively, and distrib-uted within the range of spring-born juveniles The fre-quency distribution of autumn-spawning adults showed abimodal pattern The smaller mode was similar to that ofspring-born juveniles, and the larger mode was similar tothat of autumn-born juveniles (Fig 6)

DiscussionOur results revealed apparently consistent variations inotolith growth trajectories of spring- and autumn-bornjuveniles Increment width during the early life stage wassignificantly narrower for spring-born than autumn-bornjuveniles, and the frequency distributions of the total width

of the first 5 increments were distinctive for each of thestocks Thus, the birth season of a fish can be positivelyidentified by analysis of its otolith growth trajectory Theotolith growth trajectories and frequency distributions ofthe first 5 increments of spring-spawning adults showedpatterns similar to spring-born juveniles, suggesting thatspring-spawning adults were recruited from spring-bornstock By tracking the monthly size-frequency distribution

of N taihuensis, we confirmed that the life span of thespring-born fish is 1 year [9] Thus, spring-spawning adultsare recruited from spring-born fish of the previous year

primor-dium, D-zone, L-zone, and daily increments

Number of increment 1

2 3 4 5 6

for Neosalanx taihuensis in Danjiangkou Reservoir: spring-born juveniles (filled circles, n = 79), autumn-born juveniles (open triangles, n = 68), spring-spawning adults (filled triangles, n = 101), autumn-spawning adults (open squares, n = 82), larger-sized fish in the samples of May 2008 (filled diamonds, n = 9), and larger- sized fish in the samples of January 2009 (open circles, n = 87) Error bar represents ?SD for autumn-born juveniles, autumn- spawning adults, spring-spawning adults, and larger-sized fish in the samples of January 2009, and -SD for spring-born juveniles and larger-sized fish in the samples of May 2008

Otolith growth trajectories of autumn-spawning adults

were intermediate between those of spring- and

autumn-born juveniles, and the frequency distribution of width of

the first 5 increments showed two modes, with one similar

to spring-born juveniles and the other similar to

autumn-born juveniles, indicating that autumn-spawning adults are

likely recruited from both spring- and autumn-born fish As

the autumn-spawning adults are smaller in size to

spring-spawning adults, and the cohort of the spring-born fish

disappears in July of the following year [9], we conclude

that autumn-spawning adults are recruited from the born cohort in the same year The life span of autumn-bornoffspring has not been fully tracked [9], but we concludethat at least a portion of autumn-spawning adults arerecruited from the cohort born in autumn of the previousyear Previous study has shown that the spring-spawningstock is much more abundant than the autumn-spawningstock in the population of N taihuensis in the middleYangtze [9] Thus, the putative life history of N taihuensis

spring-in Danjiangkou Reservoir can be summarized as: (1) asmall part of the spring-born fish mature and spawn inautumn of the year of their birth, (2) the majority of thespring-born cohort matures and spawns in spring of thenext year, (3) autumn-spawning adults are recruited fromboth the previous year’s autumn-born cohort and thespring-born cohort of the same year, and (4) most offspring

of the autumn-born fish mature and spawn in autumn of thefollowing year The larger-sized fish collected on both 17May 2008 and 16 January 2009 were all identified asspring-born fish, which basically agrees with the aboveproposed life history of N taihuensis The lower abun-dance of autumn-born fish compared with spring-born fish,together with the small number of larger-sized fish col-lected, might explain why no autumn-born fish wereidentified in the specimens of 17 May 2008 Similarly, theautumn-born cohort was not fully tracked in the monthlycohort-composition investigation in another middleYangtze population of N taihuensis [9] Otolith micro-structure has been successfully used for stock identification

in several other fish species, e.g., Clupea harengus L [19,20], Oncorhynchus nerka (Walbaum) [21], Gadus morhua

L [22], and Melanogrammus aeglefinus (L.) [23, 24].Our results provide reasonable cues for stock identification

of a species with a complex seasonal spawning stockstructure

Water temperature is a major factor influencing fishgrowth, particularly during early life stages [25,26] Lowerwater temperatures during the spring spawning season thanthe autumn spawning season likely contribute to the slowerearly growth of spring-born N taihuensis that we observed

in Danjiangkou Reservoir (Fig.2) While we observedcomplete separation of the early growth trajectories for fishborn in the two seasons, it is possible that annual variations

in seasonal mean water temperatures may cause this aration to be not as distinct in some years Winter mortalityplays a significant role in determining recruitment success

sep-of young-sep-of-year (YOY) in many temperate fish tions, and is often size dependent [27,28] Despite the factthat the growing season before winter is shorter forautumn-born fish than spring-born fish, rapid early growthrates allow autumn-born N taihuensis to quickly reachcritical size for overwintering survival and subsequentrecruitment to the spawning population

sagittal otoliths of Neosalanx taihuensis in Danjiangkou Reservoir

for: spring-born juveniles (a, n = 79), autumn-born juveniles (b,

n = 68), spring-spawning adults (c, n = 101), autumn-spawning

adults (d, n = 82), larger-sized fish in the samples of May 2008 (e,

n = 9), and larger-sized fish in the samples of January 2009

(f, n = 87)

Results of this study provide fundamental information

for understanding the life history of N taihuensis Multiple

seasonal spawning stocks within fish populations has been

suggested to be a reproductive strategy to reduce the

neg-ative impacts of environmental variation on reproductive

success [29,30] Our previous study on another N

taihu-ensis population in the middle reach of the Yangtze River

demonstrated that the proportion of autumn-spawning

females increased as population abundance recovered from

a low level [9] Thus, early maturation and recruitment to

spawning may be an immediate response to critically low

population levels, or fluctuating environmental conditions

such as those found in the Yangtze River Basin Further

research is necessary to illuminate the importance of

multiple maturation and spawning strategies in regulating

the population dynamics of these new but important

res-ervoir populations of N taihuensis

National Basic Research Program of China (973 Program, No.

2009CB119200), the National Natural Science Foundation of China

(No 30771642 and 30972257), the Chinese Academy of Sciences

Visiting Professorship for Senior International Scientists issued to

B.R.M (No 2010T2S25), and the One Hundred Talents Programme

of the Chinese Academy of Sciences issued to Dr S Xie and

Net-works of Aquaculture Centres in Asia and Pacific (No ICE/SL/FIS/

2007/02, NACA) Participation of B.R.M was also supported by the

Acorn Alcinda Foundation, Lewes, DE, USA.

References

1 Beacham TD, Candy JR, McIntosh B, Macconnachie C, Tabata

A, Kaukinen K, Deng L, Miller KM, Withler RE, Varnavskaya N

(2005) Estimation of stock composition and individual

identifi-cation of sockeye salmon on a Pacific rim basis using

microsat-ellite and major histocompatibility complex variation Trans Am

Fish Soc 134:1124–1146

2 Clausen LA, Bekkevold D, Hatifield EM, Mosegard H (2007)

Application and validation of otolith microstructure as a stock

identification method in mixed Atlantic herring (Clupea

haren-gus) stocks in the North Sea and western Baltic ICES J Mar Sci

64:377–385

3 Jo´nsdo´ttir IG, Marteinsdottir G, Campana SE (2007) Contribution

of different spawning components to the mixed stock fishery for

cod in Icelandic waters ICES J Mar Sci 64:1749–1759

4 Brophy D, Danilowicz BS (2002) Tracing populations of Atlantic

herring (Clupea harengus L.) in the Irish and Celtic Seas using

otolith microstructure ICES J Mar Sci 59:1305–1313

5 Dickey-Collas M, Damme CJ, Clausen LA, Fa¨ssler SM (2005)

Within stock structure and TACs: an investigation into the

spawning origin of North Sea herring using otolith microstructure

and the dynamics of Downs herring ICES CMK 12:1–14

6 Xie YH, Xie H (1997) Classification, distribution, and population

ecology of Salangidae fishes Chin J Fish 2:11–19 (in Chinese

with English abstract)

7 Hu CL, Chen WX, Liu JS (1998) Status of transplantation and

enhancement of icefishes in China and their development

strat-egies Reserv Fish 2:3–7 (in Chinese)

8 Ni Y, Zhu CD (2005) Fishes of the Taihu Lake Shanghai Scientific

and Technical Publishers, Shanghai, pp 217–221 (in Chinese)

9 Gong WB, Li HT, Xie SQ, Liu JS, Murphy BR, Xie SG (2009) Two spawning stocks of icefish Neosalanx taihuensis revealed from annual reproductive cycle analyses Fish Sci 75:1157–1165

10 Hedger RD, Atkinson PM, Thibault I, Dodson JJ (2008) A quantitative approach for classifying fish otolith strontium: cal- cium sequences into environmental histories Ecol Inf 3:207–217

11 Radhakrishnan KV, He WP, Liu M, Xie SG (2009) Otoliths—the biological CD-ROMs of fish Curr Sci 8:1121–1122

12 Perter CS, Elise RI (2008) Validation of daily ring deposition in the otoliths of age-0 channel catfish North Am J Fish Manag 28:212–218

13 Tonkin Z, King AJ, Robertson A (2008) Validation of daily increment formation and the effects of different temperatures and feeding regimes on short-term otolith growth in Australian smelt Retropinna semoni Ecol Freshw Fish 17:312–317

14 Xie SG, Watanabe Y (2005) Hatch date-dependent differences in early growth and development recorded in the otolith micro- structure of Trachurus japonicus J Fish Biol 66:1720–1734

15 Pin˜eiro C, Rey J, Pontual H, Garcı´a A (2008) Growth of west Iberian juvenile hake estimated by combining sagittal and transversal otolith microstructure analyses Fish Res 93:173–178

North-16 Fukuda N, Kuroki M, Shinoda A, Yanada Y, Okamnra A, ama J, Tsukamoto K (2009) Influence of water temperature and feeding regime on otolith growth in Anguilla japonica glass eels and elvers: does otolith growth cease at low temperatures? J Fish Biol 74:1915–1933

Aoy-17 Fu LJ, Xie YH, Li B, Zhao HC (1997) Study on daily-growth increment of otolith and the growth of larvae icefish J Fish Sci China 4:21–27 (in Chinese with English abstract)

18 Yang QR, Liu JH, Wu Q, Wang K, Zhu BK (2007) formes studies on daily-growth increment of otolith and growth

Salmoni-of Taihu icefish, Neosalanx taihuensis Freshw Fish 3:59–62 (in Chinese with English abstract)

19 Moksness E, Fossum P (1991) Distinguishing spring- and autumn-spawned herring larvae (Clupea harengus L.) by otolith microstructure ICES J Mar Sci 48:61–66

20 Stenevik EK, Fossum P, Johannessen A, Folkvord A (1996) Identification of Norwegian spring spawning herring (Clupea harengus L.) larvae from spawning grounds off western Norway applying otolith microstructure analysis Sarsia 80:285–292

21 Finn JE, Burger CV, Hollandbartels L (1997) Discrimination among populations of sockeye salmon fry with Fourier analysis

of otolith banding patterns formed during incubation Trans Am Fish Soc 126:559–578

22 Petursdottir G, Begg GA, Marteinsdottir G (2006) Discrimination between Icelandic cod (Gadus morhua L.) populations from adjacent spawning areas based on otolith growth and shape Fish Res 80:182–189

23 Begg GA, Brown RW (2000) Stock identification of haddock (Melanogrammus aeglefinus) on Georges Bank based on otolith shape analysis Trans Am Fish Soc 129:935–945

24 Begg GA, Overholtz WJ, Munroe NJ (2001) The use of internal otolith morphometrics for identification of haddock (Melano- grammus aeglefinus) stocks on Georges Bank Fish Bull 99:1–14

25 Peterson BC, Bates TD (2008) Growth and feed efficiency of juvenile channel Catfish reared at different water temperatures and fed diets containing various levels of fish meal North Am J Aquac 70:347–352

26 Cook AM, Duston J, Bradford RG (2010) Temperature and salinity effects on survival and growth of early life stage Shubenacadie River Striped Bass Trans Am Fish Soc 139: 749–757

27 Conover DO, Hurst TP (1998) Winter mortality of year Hudson River striped bass (Morone saxatilis): size-depen- dent patterns and effects on recruitment Can J Fish Aquat Sci 55:1122–1130

28 Meekan MG, Vigliola L, Hansen A, Doherty PJ, Halford A,

Carleton JH (2006) Bigger is better: size-selective mortality

throughout the life history of a fast-growing clupeid,

Spratello-ides gracilis Mar Ecol Prog Ser 317:237–244

29 Bobko SJ, Berkeley SA (2004) Maturity, ovarian cycle,

fecun-dity, and age-specific parturition of black rockfish (Sebastes

melanops) Fish Bull 102:418–429

30 Tracey SR, Lyle JM, Haddon M (2007) Reproductive biology and per-recruit analyses of striped trumpeter (Latris lineata) from Tasmania, Australia: implication for management Fish Res 84:358–367

O R I G I N A L A R T I C L E Biology

Molecular cloning of a cDNA encoding insulin-like androgenic

gland factor from the kuruma prawn Marsupenaeus japonicus

and analysis of its expression

Kota Banzai• Noriko Ishizaka• Kiyoshi Asahina•

Katsuyoshi Suitoh• Susumu Izumi•

Tsuyoshi Ohira

Received: 30 November 2010 / Accepted: 14 February 2011 / Published online: 8 March 2011

Ó The Japanese Society of Fisheries Science 2011

Abstract Androgenic gland hormone (AGH), which is

produced in the male-specific androgenic gland (AG) and

controls sex differentiation in crustaceans, has been

char-acterized only from isopod species To date,

complemen-tary DNA (cDNA) encoding an androgenic gland-specific

polypeptide, which was designated as an insulin-like

androgenic gland factor (IAG), has been cloned from three

decapod species IAG has been thought to be a candidate

for AGH in decapod crustaceans; however, there has been

no clear evidence To accumulate sequence information of

additional IAGs, we cloned cDNA encoding an IAG

pre-cursor from the kuruma prawn Marsupenaeus japonicus by

reverse-transcription polymerase chain reaction (RT-PCR)

coupled with 50- and 30-rapid amplification of cDNA ends

(RACE) The Mar japonicus IAG precursor consisted of a

signal peptide, a B chain, a C peptide, and an A chain

This organization is the same as those of the known isopod

AGH and decapod IAG precursors In RT-PCR analysis oftissue-specific gene expression of Mar japonicus IAG,transcript was detected in the distal part of medial vasdeferens, distal vas deferens, proximal part of terminalampoule, and terminal ampoule This result indicates that

AG cells of Mar japonicus are probably distributed widely

in the distal part of the male reproductive organ

Keywords Androgenic gland hormone  Crustacea Insulin-like androgenic gland factor  Kuruma prawn Marsupenaeus japonicus

IntroductionThe androgenic gland (AG) was first discovered byCharniaux-Cotton [1] in the amphipod Orchestia gamma-rella The androgenic gland produces androgenic glandhormone (AGH), which is responsible for the development

of male characteristics in crustaceans [2] Using theterrestrial isopod Armadillidium vulgare, purification ofAGH has been attempted by several research groups sincethe late 1980s [3 6] In 1999, the primary structure of

A vulgare AGH (Arv-AGH) was finally determined to be aheterodimeric glycoprotein [7], and a cDNA encodingArv-AGH precursor was cloned [8]

Pro-Arv-AGH consists of 123 amino acid residues andhas eight Cys residues The structural organization of pro-Arv-AGH, consisting of a B chain, a C peptide, and an

A chain, each linked by dibasic amino acid residues Lys–Arg, is similar to that of proinsulin, although there is nosimilarity in terms of the amino acid sequence or peptidechain length Mature Arv-AGH resulting from posttrans-lational processing by removal of the C peptide is aheterodimer consisting of the A and B chains interlinked

Course of Biological Sciences, Graduate School of Science,

Kanagawa University, 2946 Tsuchiya, Hiratsuka,

Kanagawa 259-1293, Japan

College of Bioresource Science, Nihon University,

1866 Kameino, Fujisawa, Kanagawa 252-8510, Japan

K Suitoh

Aichi Prefectural Sea Farming Institute, 1-3 Ichizenmatsu,

Konakayama, Tahara, Aichi 441-3618, Japan

Department of Biological Sciences, Faculty of Science,

Kanagawa University, 2946 Tsuchiya, Hiratsuka,

Kanagawa 259-1293, Japan

e-mail: ohirat-bio@kanagawa-u.ac.jp

Fish Sci (2011) 77:329–335

DOI 10.1007/s12562-011-0337-8

by disulfide bonds [7,9] Mature Arv-AGH has an N-linked

glycan moiety in the A chain, which was found to be

essential for biological activity [9]

In addition to Arv-AGH, cDNAs encoding AGH

precursors have been cloned from two isopods: Porcellio

scaber and P dilatatus [10] The two isopod AGH

pre-cursors consist of a signal peptide, a B chain, a C

pep-tide, and an A chain, which are organized in the same

way as in the Arv-AGH precursor, and show moderate

sequence similarities to the Arv-AGH precursor (65%

and 63%, respectively) On the other hand, cDNA

encoding AGH precursor has not yet been isolated from

any other decapod crustacean To overcome this

prob-lem, AG cDNA libraries of Cherax quadricarinatus and

Macrobrachium rosenbergii were constructed by using

suppression subtractive hybridization [11, 12] Expressed

sequence tag (EST) analysis using these cDNA libraries

revealed two AG-specific genes of C quadricarinatus

and Mac rosenbergii, respectively The two gene

prod-ucts designated as an insulin-like AG factor (IAG)

con-sist of a signal peptide, a B chain, a C peptide, and an

A chain Furthermore, the two IAG genes are specifically

expressed in the AG Therefore, the two IAG precursors

have been thought to be putative AGHs in decapod

species

The kuruma prawn Marsupenaeus japonicus is an

important aquaculture species in Japan Moreover, many

larvae and juveniles are annually produced and released

from the coasts of Japan for the purpose of sea farming

For the further development of shrimp aquaculture and

sea farming, more efficient seed production techniques

are required Therefore, it is desirable to gain better

understanding of AGH in Mar japonicus, because AGH

regulates not only male sex differentiation but also

spermatogenesis in decapod species [13] Moreover,

AGH has the potential to be utilized to control the

gender of cultured shrimps, which could be applied to

monosex shrimp culture To date, only one histological

study investigating organogenesis of AG has been

con-ducted using Mar japonicus [14], and there have been

no reports describing Mar japonicus AGH or IAG In

2010, the nucleotide sequence of a cDNA encoding the

IAG precursor from the giant tiger prawn Penaeus

monodon, which is closely related to Mar japonicus,

was revealed and deposited in the GenBank database

(accession no GU208677) However, the sequence has

not been formally published yet In this study, we

iso-lated a cDNA encoding Mar japonicus IAG (Maj-IAG)

precursor by RT-PCR coupled with 50- and 30-RACE

using primers designed based on the nucleotide sequence

of P monodon IAG (Pem-IAG) cDNA Subsequently, we

examined the tissue-specific gene expression of Maj-IAG

Reverse-transcription PCRTotal RNA from distal vas deferens of Mar japonicus wasprepared by the acid guanidium thiocyanate–phenol–chlo-roform extraction method [17] First-strand cDNA wassynthesized with 500 ng total RNA using SuperScript IITMreverse transcriptase (Invitrogen, Carlsbad, CA, USA)according to the manufacturer’s protocol In this reaction,NotI dT primer (GE Healthcare UK, Buckinghamshire,

japonicus T testis, PVD proximal vas deferens, MVD medial vas deferens, DMVD distal part of medial vas deferens, DVD distal vas deferens, PTA proximal part of terminal ampoule, TA terminal ampoule

England) containing the nucleotide sequences of an

oligo-dT and 30-adaptor primers (RTG and RTG-NN in Table1)

was used as a reverse-transcription primer For RT–PCR,

two oligonucleotide primers (F and R in Table1) were

designed based on the nucleotide sequences of P monodon

insulin-like androgenic gland factor (Pem-IAG) precursor

(accession no GU208677) The first-strand cDNA was

used as a template, and amplification was primed by a pair

of the primers (F and R; Fig.2) The following program

was used for PCR amplification: 35 cycles of 30 s at 94°C

(3 min 30 s for the first cycle only), 30 s at 55°C, and

1 min at 72°C (3 min for the last cycle only)

30-RACE

Two specific oligonucleotide primers (3F and 3NF in

Table1) were designed based on the nucleotide sequence

of the Maj-IAG cDNA fragment amplified by RT-PCR A

cDNA fragment encoding the 30-region of Maj-IAG was

amplified by two rounds of PCR In the first PCR, the strand cDNA described in RT-PCR was used as a template,and amplification was primed by a pair of the primers, 3Fand RTG, as shown in Fig.2 In the second PCR, the firstPCR product was used as template, and amplification wasprimed by a pair of nested primers, 3NF and RTG-NN, asshown in Fig.2 The following program was used for eachPCR amplification: 35 cycles of 30 s at 94°C (3 min 30 sfor the first cycle only), 30 s at 55°C, and 3 min at 72°C(6 min for the last cycle only)

first-50-RACEFirst-strand cDNA was newly synthesized with 500 ngtotal RNA from distal vas deferens by using a SMARTRACE cDNA amplification kit (Clontech, Mountain View,

CA, USA) according to the user’s manual except thatMoloney murine leukemia virus (MMLV) reverse trans-criptase (Clontech) was substituted by SuperScript IITMreverse transcriptase (Invitrogen) In this reaction, a

50-cDNA synthesis (CDS) primer (Clontech) was used as areverse-transcription primer For 50-RACE, two specificprimers (5R and 5NR in Table1) were designed based onthe nucleotide sequence of the Maj-IAG cDNA fragmentobtained from RT-PCR A cDNA fragment encoding the

50-region of the Maj-IAG cDNA was amplified by tworounds of PCR In the first PCR, the newly synthesizedfirst-strand cDNA was used as a template, and amplifica-tion was primed by a pair of primers, 5R and UniversalPrimer Mix (UPM, Clontech), as shown schematically inFig.2 In the second PCR, the first PCR product was used

as a template, and amplification was primed by a pair ofnested primers, 5NR and Nested Universal Primer (NUP,Clontech), as in Fig 2 The following program was usedfor each PCR amplification: 5 cycles of 25 s at 94°C(3 min 25 s for the first cycle only) and 3 min at 72°C, 5cycles of 25 s at 94°C, 10 s at 68°C (70°C in the second

Marsupenaeus japonicus IAG (Maj-IAG) and the locations of the

oligonucleotide primers Arrowheads represent the primers, and

lines under the arrowheads indicate the amplified cDNA fragments Boxes represent the open reading frame Lines on both sides of the