Aquaculture nutrition, tập 19, số 1s, 2013

Aquaculture Nutrition 2013 19; 1–24 doi: 10.1111/anu.12091 2 Aquaculture and Fisheries Group, Wageningen University, Wageningen, The Netherlands; Section for Coastal Ecology, DTU Aqua, National Institute of Aquatic Resources, Danish Technical University, Charlottenlund, Denmark; INRA, UR1067 Nutrition Metabolism Aquaculture, Saint-P ee-sur-Nivelle, France; National Institute of Sciences and Technolo5 gies of the Sea, Tunis, Tunisia; Biomar A/S, Brande, Denmark To sustain eel aquaculture, development of reproduction in captivity is vital The aim of this review is to assess our current knowledge on the nutrition of broodstock eels in order to improve the quality of broodstock under farming conditions, drawing information from wild adult eels and other marine pelagic spawners Freshwater eels spawn marine pelagic eggs with an oil droplet (type II), and with a large perivitelline space Compared with other marine fish eggs, eel eggs are at the extreme end of the spectrum in terms of egg composition, even within this type II group Eel eggs contain a large amount of total lipids, and a shortage of neutral lipids has been implied a cause for reduced survival of larvae Eel eggs have higher ARA but lower EPA and DHA levels than in other fish Too high levels of ARA negatively affected reproduction in the Japanese eel, although high levels of 18:2n-6 in the eggs of farmed eels were not detrimental The total free amino acid amount and profile of eel eggs appears much different from other marine pelagic spawners Nutritional intervention to influence egg composition seems feasible, but responsiveness of farmed eels to induced maturation might also require environmental manipulation The challenge remains to succeed in raising European eel broodstock with formulated feeds and to enable the procurement of viable eggs and larvae, once adequate protocols for induced maturation have been developed KEY WORDS: amino acids, Anguilla spp., broodstock nutrition, fatty acids, feed, minerals, nutrients, vitamins ª 2013 John Wiley & Sons Ltd Received 26 August 2012; accepted May 2013 Correspondence: L Heinsbroek, Aquaculture and Fisheries Group, Wageningen University, Wageningen 6700 AH, The Netherlands E-mail: leon.heinsbroek@wur.nl Recruitment and wild stock of European eel (Anguilla anguilla, L.) have declined drastically over the last decades Habitat reduction and over-fishing, climate change, pollution and infections with the swim bladder parasite (Anguillicoloides crassus) and/or eel viruses have been implicated as causes for the current decline of the eel population (van Ginneken & Maes 2005) The major part of eel production now comes from aquaculture, but this is still capture based, relying on wild caught glass eels To sustain eel aquaculture, development of reproduction in captivity is vital Research on eel reproduction is complicated, because broodstock eels stop feeding when silvering in nature Although silvering is reversible and feeding can be resumed when migration is not initiated (Sved€ ang & Wickstr€ om 1997; Durif & Elie 2008), it has been shown for A japonica that eels caught in the spawning area had not been eating in the marine phase of the migration (Chow et al 2010) Also in captivity, feeding is terminated after transfer to saltwater prior to induction of maturation Thus, for eels, all the qualitative and quantitative requirements for reproduction have to be met from their body reserves highlighting the importance of prespawning nutrition For eel embryos and larvae, the expression ‘you are what you eat’ might be extended to ‘you are what your parents ate a long time ago’ Furthermore, the life history of A anguilla, being the latest in the anguillid evolution (Aoyama 2009), is in a number of respects also the most extreme They have the longest migration distance, the longest larval duration and the highest body lipid levels and are least mature even in the silver stage Especially in A anguilla, silver females puberty, defined as the onset of vitellogenesis (Taranger et al 2010), is not yet started (Durif et al 2006; Palstra et al 2011) Other temperate eels, A japonica (Sudo et al 2011a), A rostrata (Cottrill et al 2001), and A australis and A dieffenbachii (Todd 1981; Lokman et al 1998), seem to be more advanced To close the life cycle of the European eel, information on larval, juvenile and adult (broodstock) nutrition is required The aim of this review is to assess our current knowledge on broodstock nutrition or on nutritional influences on reproduction of A anguilla, in order to improve the quality of broodstock under farming conditions In nature, a large part of the reproductive investment of anguillid eels is spent during migration The energy requirements during migration also consist of a ‘fixed’ part (standard metabolic rate) and a ‘variable’ part (active metabolism above standard) Therefore, the total costs of transport (COT, kJ kgÀ1 kmÀ1) are influenced by both the distance to the spawning area and the swimming speed (Palstra & van den Thillart 2010) At optimal swimming speeds of 0.4–0.6 m sÀ1, both females and males eels have a COT of 0.4–0.7 kJ kgÀ1 kmÀ1, when measured by oxygen consumption, or a COT of 0.6–1 kJ kgÀ1 kmÀ1, when estimated from the body energy losses (Van Ginneken et al 2005a; Palstra et al 2008; Burgerhout et al 2010) Based on this, and depending on the initial body energy content, A anguilla uses between 15% and 40% of its initial energy reserves for migration Of this energy, about 70–80% is provided by body lipids, the remainders mostly by body protein (Bo€etius & Bo€etius 1980, 1985; Van Ginneken et al 2005a) Energy (and nutrients) invested during gonad development is either deposited in the gonads or used to ‘fuel’ this deposition When artificially matured, A anguilla males reach gonado-somatic indices (GSI) of 6–14% (Bo€etius & Bo€etius 1967; Amin 1991; Van Ginneken et al 2005b; Mazzeo et al 2010) For A japonica males, GSI of up to 40% have been reported (Lau 1987; Tsukamoto et al 2011), but it is not clear whether this is a true difference or the result of a further advanced emaciation of these eel’s soma Maturing female eels reach GSI of 40–60% or, due to variable degree of hydration (Fig 5), 22–28% on a dry matter (dm dmÀ1) basis (Bo€etius & Bo€etius 1980; Lau 1987; Yamada et al 2001) Little is known about the composition of the testes and the semen of eels Cheung (1983) and Lau (1987) reported for A japonica testes lipid levels of 8–11 g kgÀ1 (30–60 g kgÀ1 dm), for both immature and ripe testes Testes lipid levels were a factor 10 lower than ovary lipid levels The few data available for A anguilla suggest that testes lipid levels are higher in this species, 90–160 g kgÀ1 (430–540 g kgÀ1 dm), with little variation between immature and ripe testes (Kokhnenko et al 1977; Amin 1991) Indirect evidence for a high lipid level of the immature testes of A anguilla can also be deduced from the testes fatty acid profile (Mazzeo et al 2010); Table 3) Due to the high lipid level, testes protein levels are a little bit lower in A anguilla, but total energy deposited in the testes seems similar between species, 0.8–1 MJ kgÀ1 initial body mass or 5–10% of the initial body energy (Lau 1987; Amin 1991) Lipid levels in the immature and early stage ovaries of A anguilla are higher (150–300 g kgÀ1 or 500– 650 g kgÀ1 dm, Palstra et al 2006; Bo€etius & Bo€etius 1980; Amin 1991; Kokhnenko et al 1977; Mazzeo et al 2011) than of A japonica (50–150 g kgÀ1 or 300– 500 g kgÀ1 dm (Lau 1987; Cheung 1983; Ozaki et al 2008)) In later stages, ovary lipid levels are comparable between species, 50–60 g kgÀ1 or 400-500 g kgÀ1 dm (Bo€etius & Bo€etius 1980; Ozaki et al 2008) Still, also egg lipid levels in A anguilla seem to be somewhat higher, 60–65 g kgÀ1 (Corraze et al 2011) versus 42–55 g kgÀ1 for A japonica (Furuita et al 2006; Tanaka et al 2006; Ozaki et al 2008) There are no indications that the protein levels in the ovaries and the eggs differ between species Deposition of total lipid and (crude) protein in the ovaries of A anguilla (Bo€etius & Bo€etius 1980) is shown in Fig Initially more lipid is deposited, but in mature ovaries, both lipid and protein converge at c 50 g kgÀ1 initial body mass Palstra et al (2006) reported a lipid deposition of 57 Æ 22 g kgÀ1 Based on these lipid and protein depositions, energy deposition in the ovary of A anguilla is estimated to be 3–3.5 MJ kgÀ1 initial body mass, or 17–23% Aquaculture Nutrition 19; 1–24 ª 2013 John Wiley & Sons Ltd Ovary protein (g/kg0.8/d) 50 40 30 20 (b) y = 0.68x – 0.25 R² = 0.71 0.8 0.6 0.4 0.2 0.0 0.0 10 0.5 1.0 1.5 2.0 Protein use (g/kg0.8/d) 10 20 30 40 50 Ovary energy (kJ/kg0.8/d) Protein, lipid (g/kg initial body mass) 1.0 (a) 60 GSI (%) (c) 60 y = 0.37x – 1.04 R² = 0.43 50 40 30 20 10 0 50 100 150 Energy use (kJ/kg0.8/d) Figure Lipid and (crude) protein deposition in the ovary of artificially matured wild silver A anguilla Calculated from Bo€etius & Bo€etius (1980), who matured a series of 27 female eels, and afterwards determined the mass and the composition of the ovaries and the remaining soma Data on initial ovary composition from Palstra et al (2006, 2011) of the initial body energy, which again seems somewhat higher than for A japonica Gonad development, or deposition of mass and energy in the gonads, is not 100% efficient and in itself costs energy From the initial body mass and composition and the final mass (soma plus ovary) and composition, the costs of deposition can be determined (Fig 1) Protein deposition is quite efficient, 68%, while energy efficiency is lower, 37%, although the latter could be determined with less certainty, probably caused by uncertainty about the individual initial lipid contents of these wild eels (Bo€etius & Bo€etius 1980) With this energy efficiency, the total energy requirement for ovary development would become 8–9.5 MJ kgÀ1 initial body mass, or 46–62% of the initial body energy Marine pelagic fish eggs In marine pelagic fish types are recognized: eggs without and eggs with droplet(s), the latter being the most common, in in temperate and warmwater species (Ahlstrom eggs, two visible oil particular & Moser Aquaculture Nutrition 19; 1–24 ª 2013 John Wiley & Sons Ltd 1980) Eels spawn marine pelagic eggs with an oil droplet, but also with a large perivitelline space, which is less common (Ahlstrom & Moser 1980; Tsukamoto et al 2011) Rønnestad et al (1999) showed that eggs with oil droplets (by them classified as type II) differed in composition but also in embryonic metabolism The type II eggs contain more lipids, and within these lipids, a (much) larger fraction is neutral lipids (Table 1) In the neutral lipids, they further contain a larger (although variable) fraction of wax- and sterol-esters (Wiegand 1996) All marine pelagic eggs contain similar amounts of total amino acids, but these are more present as free amino acids (FAA) in the type I eggs Apart from the role of FAA in early embryonic energy metabolism (section Protein and amino acids), they also function as osmotic effectors in the acquisition of egg buoyancy (Rønnestad et al 1999; Cerd a et al 2007; Finn & Fyhn 2010) Although eel eggs can be categorized as type II eggs, even within this group, eel eggs are at the extreme end of the spectrum (Table 1) If and how this position, which is extended in the fatty acid and FAA profiles (sections Lipids and Protein and amino acids), is related to the large perivitelline space is not clear (Unuma et al 2005) Table Size and composition of marine pelagic fish eggs Other marine pelagic Egg diameter (mm) Oil glob diameter Egg dm ug eggÀ1 Moisture (g kgÀ1) Total lipids (g kgÀ1 dm) NL (%TL) TG (%NL) CH (%NL) WSE (%NL) PL (%TL) PC (%PL) PE (%PL) PI (%PL) Total N (g kgÀ1 dm) Protein (g kgÀ1 dm) FAA (g kgÀ1 dm) Carbohydrates (g kgÀ1 dm) Ash (g kgÀ1 dm) Anguilla spp.1 Oil globule2 No oil globule3 1.1–1.8 0.25–0.35 60–70 880–920 350–440 0.8–1.5 (5) 0.15–0.4 35–120 880–930 130–450 0.8–2 (6) – 150 900–930 70–150 40–84 30–84 6–20 6–65 16–60 60–86 12–20 2–6 90–110 250–500 100–170 3–24 25–40 25–45 20–50 6–15 60–75 65–88 7–25 1–2 100–120 350–550 150–220 6–20 60–100 60–150 80–83 30–40 17–20 95 300–460 30–50 83 €tius & Boe €tius Data on A anguilla (Kokhnenko et al 1977; Boe 1980; Bezdenezhnykh & Prokhorchik 1984; Prokhorchik 1987; Pedersen 2004; Palstra et al 2005; Corraze et al 2011), A rostrata (Edel 1975; Oliveira & Hable 2010), A australis (Lokman & Young 2000) and A japonica (Seoka et al 2003, 2004; Unuma et al 2005; Furuita et al 2006, 2007; Tanaka et al 2006; Ohkubo et al 2008; Ozaki et al 2008; Kagawa et al 2009) Data on Dicentrarchus labrax (Devauchelle & Coves 1988;  et al 1994a; Bell et al 1997; Navas et al 1997, 2001; Cerda Rønnestad et al 1998b), Sparus auratus (Mourente & Odriozola ndez-Palacios et al 1995, 1990; Rønnestad et al 1994; Ferna 1997; Rodrıguez et al 1998; Almansa et al 1999, 2001), Pagrus major (Watanabe et al 1984c, 1985b; Seoka et al 1997), Dentex dentex (Tulli & Tibaldi 1997; Mourente et al 1999; nez et al 2008; Samaee et al 2009a,b, 2010), Diplodus sarGime rez et al 2007), Scophthalmus maximus gus (Cejas et al 2003; Pe (McEvoy et al 1993; Rainuzzo et al 1994; Silversand et al 1996), Scophthalmus rhombus (Cruzado et al 2011), Paralichthys olivaceus (Furuita et al 2000, 2002, 2003c), Seriola quinqueradiata (Verakunpiriya et al 1996), Seriola lalandi (Moran et al 2007; Hilton et al 2008), Sciaenops ocellata (Vetter et al 1983), Latris lineata (Morehead et al 2001; Brown et al 2005), Pseudocaranx dentex (Vassallo-Agius et al 1998, 2001a), Lates calcarifer (Southgate et al 1994; Sivaloganathan et al 1998; Dayal et al 2003) and Rachycentron canadum (Faulk & Holt 2003, 2008; Nguyen et al 2010, 2012) Data on Gadidae, Gadus morhua (Craik & Harvey 1984; Fraser et al 1988; Finn et al 1995a,b; Salze et al 2005; Penney et al 2006) Melanogrammus aeglefinus (Craik & Harvey 1984; Reith et al 2001) Theragra chalcogramma (Ohkubo et al 2006) and Pleuronectidae, Hippoglossus hippoglossus (Falk-Petersen et al 1986, 1989; Rainuzzo et al 1992; Bruce et al 1993; Evans et al 1996; Mazorra et al 2003) Pleuronectes platessa (Craik & Harvey 1984; Rainuzzo et al 1992; Thorsen & Fyhn 1996), Microstomus kitt (Thorsen & Fyhn 1996) and Verasper moseri (Ohkubo & Matsubara 2002) Lipids In most instances, reproduction of wild fish, or of fish fed with natural food, is more successful than of farmed fish This has also been shown for A anguilla (Tomkiewicz 2012), but less so for A japonica, at least when feminized eels are used (Yamada et al 2006) One of the differences between wild and farmed eels (and between males and females) are the body lipid levels (Fig 2) In A anguilla, males start to silver, that is, initiating maturity, at 80–150 g, while for females this is at sizes from 300 to 800 g (Tesch 2003) Silvering eels, farmed and wild eels of both sexes, attain lipid levels of 25–35% [a.o (Bo€etius & Bo€etius 1985; Larsson et al 1990; Kamstra & Van Heeswijk 1996; Garcia-Gallego & Akharbach 1998; Kn€ osche 2009; Clevestam et al 2011)] Both farmed and wild males reach these levels early, at masses of 70–100 g Wild female eels seem to follow another trajectory: they first invest in body mass growth and reach these higher lipid levels at higher body masses (Fig 2) Although most farmed females also have lower body lipid levels than males of comparable mass (Kamstra & Van Heeswijk 1996), their trajectory is clearly advanced compared with wild females Anguilla anguilla silver eels generally have much higher body lipid levels than other temperate eels (150–230 g kgÀ1, Han et al 2000, 2001; Tremblay 2009; De Silva et al 2002; Hopkirk et al 1975) In vertebrates, the body lipid mass, through adiposity signals leptin and insulin, is thought to influence reproduction in a number of ways (Caprio et al 2001) On the one hand, a minimum lipid mass seems to be required to initiate puberty Such an effect was also observed in Oncorhynchus mykiss (Weil et al 2008) Furthermore, Peyon et al (2001) showed in Dicentrarchus labrax that (recombinant mouse) leptin stimulated in vitro pituitary LH release in the prepubertal stage, but much less in later stages For A anguilla, Larsson et al (1990) even hypothesized that body lipid content might be the trigger for silvering, but as silver A anguilla are prepubertal and also some silver eels have (very) low body lipid levels (Sved€ ang & Wickstr€ om 1997), this seems not so On the other hand, excessive body lipid stores negatively affect reproduction, through impairment of gonadal steroidgenesis (Caprio et al 2001) Evidence for a negative effect of increased adiposity on reproduction in fish is mostly anecdotal, in Indian (Chaudhuri, 1960) and Chinese carps (Chen et al 1969, cited by Rath et al 1999) However, the observed negative effects of a low protein broodstock diet in Dicentrarchus labrax (Cerd a et al 1994b) could well also have originated from a lower DP/DE ratio, as the gonads (but not the eggs) of the deficient fish showed higher lipid levels during peak spawning Finally, Aquaculture Nutrition 19; 1–24 ª 2013 John Wiley & Sons Ltd (a) Body/Muscle lipid (g/kg) 400 350 300 250 200 150 100 50 0 200 400 600 800 1000 1200 1400 1600 1200 1400 1600 Body mass (g) (b) 400 Body/Muscle lipid (g/kg) Figure Whole body or muscle lipid percentage of wild (a) and farmed (b) A anguilla in relation to body mass, and sex (females = circles; males = triangles) The light grey symbols in (a) are for yellow eels, and the dark grey symbols are for silver eels Data on wild eels are from the study by Bo€etius & Bo€etius (1985, 1989) and Larsson et al (1990) and Heinsbroek, unpublished, IMARES, unpublished Data on farmed eels are from the study by Kamstra & Van Heeswijk (1996), Schmitz (1982), Corraze et al (2011), Støttrup et al (2013) and Heinsbroek, unpublished 350 300 250 200 150 100 50 0 the link between adiposity signals and the dopaminergic system (Baskin et al 1999) might explain (partly) the difference in responsiveness to maturation between farmed and wild eels and the fact that for A anguilla, substantially more weekly injections are needed, 12–25 (Durif et al 2006; Palstra & Thillart 2009), to complete maturation than for other eels, 8–14 injections (Lokman & Young 2000; Kagawa et al 2005; Oliveira & Hable 2010) Based on the above, one might have assumed that a decrease in body lipids during starvation and swimming could be stimulating for reproduction However, Bo€etius & Bo€etius (1985) and Van Ginneken et al (2005a) showed that starvation had no effect on body composition, indicating that energy use from lipid and protein was in the same proportion as in the body composition Van Ginneken et al (2005a) found that even prolonged swimming had no effect on the body composition of A anguilla This may be Aquaculture Nutrition 19; 1–24 ª 2013 John Wiley & Sons Ltd 200 400 600 800 1000 Body mass (g) due to the fact that they did used farmed eels with a very high body lipid content (340 g kgÀ1) It might be that at lower body lipid levels, a decrease does occur, as indicated by the results of Larsson & Lewander (1973) and Dave et al (1975) who reported a decline in muscle lipids from 90 to 30 g kgÀ1 in yellow A anguilla fasted for more than months at 2–10 °C Induced maturation of A anguilla is reported to induce no (Palstra et al 2006) or only a slight (Mazzeo et al 2011) decrease in the muscle lipid content Ozaki et al (2008) also reported no change in lipid content in the muscle of A japonica during induced maturation; however, Lau (1987) and Liu et al (2009) did find a strong decrease In insects and birds, it has been shown that the demands for specific lipid classes and FA differ between migration and reproduction (Zhao & Zera 2002) Sasaki et al (1989) also found a change in lipid class and FA composition in the muscle of migrating Oncorhynchus keta As further specific FA (n-6 and n-9) are implicated in swimming capacity (McKenzie et al 1998; Chatelier et al 2006), it might be that FA are selectively allocated to migration and reproduction and that if some FA are not used for migration, they might negatively influence the egg composition Liu et al (2009) did find in A japonica that fasting and swimming led to a selective retention of ARA in the muscle This was not the case in eels that were induced to maturate, indicating a selective incorporation in the ovary Total lipid levels in eggs of A japonica are normally reported to be 300–450 g kgÀ1 dm (Furuita et al 2003a, 2006; Tanaka et al 2006) although Unuma et al (2005) also mention TL levels as low as 200 g kgÀ1 dm Furuita et al (2006) did find a negative correlation between egg TL and fertilization, hatching and survival Surprisingly this effect was mainly caused by higher levels of PL A similar effect was also described in Hippoglossus hippoglossus by Evans et al (1996), but for the relative amount of PL These authors therefore suggested that this was more an indication of a lack of NL Total lipid levels remain stable 350–400 g kgÀ1 dm in A japonica eggs until hatching and decrease during yolk sac and oil droplet resorption to c 160 g kgÀ1 dm (Tanaka et al 2006) Ohkubo et al (2008) showed that during this period, TG decreased stronger (80%) than PL (40%) A shortage of neutral lipids has been implied as a cause for larval mortality in Seriola lalandi (Hilton et al 2008) and Latris lineata (Morehead et al 2001) This might well also have contributed to mortalities before first feeding, or even mouth formation, in larvae of A australis (Lokman & Young 2000) and A rostrata (Oliveira & Hable 2010) Another striking difference between wild and farmed fish lies in the fatty acid profile of the eggs Clear relations between FA profile and egg quality have however not always been apparent (Fernandez-Palacios et al 2011) The egg fatty acid profile of wild A japonica is compared with other marine pelagic spawners, for both type I and II, in Table The differences can partly be explained by differences in lipid class composition (Table 1), but again eels are at the far end, or even outside, the spectrum of type II eggs A japonica eggs have lower levels of EPA and especially of DHA, and much higher levels of 18 : 1, also in the PL The lipid class profile of eel eggs (Table 1) suggests that the majority of the PL in the eggs originate from vitellogenin, as also shown in other fish (Silversand & Haux 1995; Johnson 2009)] Vitellogenin (VTG) of A japonica was actually one of the first teleost VTGs characterized (Hara Table Fatty acid profile (% of total FA) of egg total and polar lipids of wild (or fed with raw fish and/or squid) A japonica and of other marine pelagic spawners Other marine pelagic Total lipids 16:0 18:1 18:2 n-6 18:3 n-3 ARA EPA 22:5n-3 DHA EPA/ARA DHA/EPA Polar lipids 16:0 18:1 18:2 n-6 18:3 n-3 ARA EPA 22:5n-3 DHA EPA/ARA DHA/EPA A japonica1 Oil globule2 No oil globule3 18.9 33.5 0.9 2.1 2.9 2.4 8.9 2.0 3.4 18.9 (13–21.5) 19.2 (9.7–25.7) 1.8 (0.3–7.5) 0.6 (0.1–1.1) (0.5–3.7) 6.4 (2.4–11) (1–4.6) 24.1 (13.7–31.4) 3.7 (0.6–8.6) 4.1 (2–6.8) 20 (17.3–23.5) 14.8 (11.2–17.6) (0.3–2.7) 0.3 (0.2–0.5) 1.9 (1–3) 13.4 (8.7–15.5) 1.5 (1.2–1.8) 28.8 (25.5–31.1) 7.8 (4.4–14) 2.2 (1.8–2.9) 21.0 11.7 0.7 0.3 3.6 9.1 2.4 32.5 3.1 3.6 21.4 13.2 0.6 0.2 2.6 13.0 1.0 32.2 5.4 2.5 (18.1–21.6) (31.4–35) (0.9–3.8) (2.1–3.7) (6.1–12) (0.7–3.3) (2–4.1) 21.2 (18.9–23.1) 22.3 (15.4–24.2) 3.8 (2.5–5) 5.3 (3.7–6.4) 17.2 (13.6–21.4) 1.4 (0.5–2.2) 3.3 (2.4–3.9) (18.5–24.1) (10.7–13) (0.1–1.9) (0.1–0.5) (1.8–4.9) (6.8–10.1) (1.2–4.6) (27–37.3) (1.4–5.4) (3.2–4) (20.9–22) (11.1–14.3) (0.03–0.9) (0.01–0.4) (1.5–3.3) (10.9–15) (0.5–1.4) (29.3–34.8) (3.9–8.8) (2.2–2.9) Furuita et al (2003a) and Ozaki et al (2008) Data on Dicentrarchus labrax (Bruce et al 1999; Navas et al 2001), Sparus aurata (Mourente & Odriozola 1990), Scophthalmus maximus (Peleteiro et al 1995; Silversand et al 1996; Lavens et al 1999), Rachycentron canadum (Faulk & Holt 2003, 2008; Nguyen et al 2010, 2012), Plectorhynchus cinctus (Li et al 2005) Centropomus undecimalis (Yanes-Roca et al 2009), Pseudocaranx dentex (Vassallo-Agius et al 1998, 2001a), Lutjanus campechanus (Papanikos et al 2008), Coryphaena hippurus (Divakaran & Ostrowski 1989; Ostrowski & Divakaran 1989), Solea senegalensis zquez 1996), Solea solea (Lund et al 2008), Para(Mourente & Va lichthys adspersus (Wilson 2009) and Centropristis striata (Bentley et al 2009) Data on Hippoglossus hippoglossus (Falk-Petersen et al 1989; Bruce et al 1993; Mazorra et al 2003) and Gadus morhua (Fraser et al 1988; Pickova et al 1997; Salze et al 2005; Penney et al 2006; Lanes et al 2012) et al 1980) It is a very specialized high-density serum phospho-lipo-glyco-protein consisting 830–860 g kgÀ1 protein and 130–170 g kgÀ1 lipid PL and TG account for 650–720 and 170–270 g kgÀ1, respectively, of the lipids (Ando & Matsuzaki 1996; Komatsu et al 1996) The fatty acid profile of eel VTG is not known, but Silversand & Haux (1995) showed for a number of fish species that the fatty acid profile of VTG and the egg PL were highly correlated They did find species-specific differences in VTG FA profiles In general, egg PL FA seem to be less affected by the broodstock diet (Mourente & Odriozola 1990; Aquaculture Nutrition 19; 1–24 ª 2013 John Wiley & Sons Ltd of 10.4%FA Similarly high levels of ARA (4-20%FA) were observed in ovaries and testes of a number of tropical reef species (Ogata et al 2004; Suloma & Ogata 2011) A higher level of ARA in the PL of the (immature) ovaries than in the eggs of A japonica was also observed by Furuita et al (2007) Remarkably, and contrary to the findings in other marine fish, in A japonica, the egg ARA can be formed from conversion of dietary 18:2 (Yamada et al 2006; Furuita et al 2007; Ozaki et al 2008) The capacity of anguillid eels to elongate and desaturate FA is well documented (Takeuchi et al 1980; Kissil et al 1987), so although of marine origin, they truly earn the name of freshwater eels (NRC 2011) This capacity is also reflected in the egg composition of A japonica Yamada et al (2006) showed that A japonica fed sunflower oil, rich in 18:2, produced eggs with twice as much ARA than eels fed with a marine oil (Fig 5) A similar but less dramatic effect of dietary 18:2 on egg ARA levels was observed by Furuita et al (2007) and, although less pronounced, by Ozaki et al (2008) A selective incorporation of FA was also observed in the testes and semen of farmed A anguilla (Mazzeo et al 2010) (Table 3) Whereas the FA profile of the immature testes is essentially the same as that of the muscle, in the mature testes, the levels of EPA, DHA and especially ARA are increased Remarkably, in the semen, EPA and ARA are further increased, but not DHA Although Perez et al reported much lower levels of LC-PUFA in the semen of farmed A anguilla, they also found a low DHA/EPA ratio which seems to be unique among the few marine teleosts studied, Dicentrarchus labrax (Bell et al 1996; Asturiano Wiegand 1996), although Silversand et al (1995) did find higher levels of 18:2 and lower levels of EPA in both VTG and eggs of farmed Gadus morhua Increased incorporation of 18:2 was also noted in the egg PL (and NL) of farmed Scophthalmus maximus (Silversand et al 1996) and Dicentrarchus labrax (Bell et al 1997) DHA is highly conserved and selectively incorporated in the PL (Table 2), suggesting the importance of DHA for embryonic and larval development (Sargent 1995; Wiegand 1996) Selective incorporation of DHA is also seen in A japonica, be it at a lower level than for other fish (Furuita et al 2007; Ozaki et al 2008) Despite the selective incorporation of DHA into the gonads and the eggs, both low levels of DHA and imbalanced LC-PUFA ratios in the broodstock diets can lead to lower DHA levels in the egg lipids (Bell et al 1997; Almansa et al 1999; Bruce et al 1999) However, no effect of broodstock diet on the DHA content of the eggs was found in A japonica (Furuita et al 2007; Ozaki et al 2008)) In both studies, the EPA level in the eggs decreased with the replacement of fish oil by corn oil in the broodstock diets, similarly to the results of Yamada et al (2006) with sunflower oil (Fig 5), and those in Gadus morhua (Silversand et al 1995) In most marine fish species, ARA is also selectively incorporated, but even more in the gonad than in eggs Perez et al (2007) found high levels of ARA accumulated in gonad PL of male and female Diplodus sargus and selective retention of this fatty acid after gonad recession There seem to be large species differences, however (Table 2) In Lutjanus argentimaculatus, Emata et al (2003) reported for eggs an already low EPA/ARA ratio of 0.9; in the ovary this ratio was only 0.2, with an ARA level Table Fatty acid profile (% of total FA) of muscle, liver, testes and semen of farmed A anguilla before and after induction of maturation1 Week2 Tissue Fatty acid 16:0 18:1 18:2n-6 18:3n-3 20:1 ARA EPA 22:5n-3 DHA EPA/ARA DHA/EPA 7–11 Muscle Liver Testis Muscle Liver Testis Semen 18 28.9 7.2 1.6 8.9 0.6 3.7 7.8 6.2 2.1 21.7 25.3 5.3 0.7 5.3 1.5 4.4 2.4 14 2.9 3.2 18.1 29.2 6.7 1.1 9.1 0.7 3.6 2.1 7.7 5.1 2.1 18.3 31 7.1 0.9 8.5 0.7 3.2 2.1 7.6 4.6 2.4 24.5 24.7 6.5 0.9 4.2 1.1 4.5 2.2 15 4.1 3.3 19.8 21.5 4.2 0.3 4.5 8.5 2.2 19.1 1.9 2.2 20 14 3.3 rez et al (2000) and 7–11 weeks: Mazzeo et al (2010) and 5–13 weeks: Pe In both studies, eels were weekly injected with HCG and started spermiating in week Aquaculture Nutrition 19; 1–24 ª 2013 John Wiley & Sons Ltd 3.6 16 1.6 22 2.3 1.4 5–13 Semen 28.2 13.4 2.5 0.2 2.2 5.2 8.5 0.8 7.7 1.6 0.9 et al 2001) and Seriola quinqeradiata (Verakunpiriya et al 1996) It is not known whether there are differences between semen of wild and farmed A anguilla as shown in Dicentrarchus labrax (Bell et al 1996; Asturiano et al 2001), nor if and how the fatty acid profile of the semen affects fertilization and embryonic development In eels, most of the egg NL seem to be transported by VLDL Endo et al (2011) showed that already during previtellogenesis, in the lipid droplet stage, both 11-KT and VLDL were required to stimulate oocyte growth and lipid incorporation Furthermore, Ando & Matsuzaki (1996) found that the plasma lipoproteins of A japonica were dominated by VLDL (25–30 g LÀ1 or more than 40% of total lipoproteins), even after induction of vitellogenesis with E2 injections, which brought VTG to 20 g LÀ1 Next to deposition of transported lipids, a substantial amount of NL in the ovary of A anguilla originates from de novo lipid synthesis within the ovary Bo€etius et al (1991) monitored the incorporation of radioactivity from 14C-acetate during a 24-h period after injection in male and female A anguilla in different stages of maturation They found that at early stages of maturation, gonad lipid synthesis equalled that in the liver For females, this coincided with the period of maximum lipid deposition in the ovaries, at GSI of 5-13% (cf Fig 1) In this period, radioactivity was mainly incorporated in TG, with 16:0 and 18:0 as major FA In later stages of ovary development, also sterol esters became important as well as monoenes and FA with more than 18 carbon atoms Bo€etius et al (1991) did find only minor synthesis of wax esters in the ovaries This might be due to fatty alcohols being synthesized as FA (not necessary de novo) in the liver and only transformed to alcohols after transport to the ovary, as described by Bell et al (1997) However, the low DHA level in the neutral lipids of anguillid eggs is another indication that wax esters are not abundant, because in other type II eggs, it was shown that the fatty alcohols were mainly saturated (mainly 16:0), or monoenes (mainly 18:1), but the FA consist for almost half of n-3 LC-PUFA, of which 50-70% DHA (Joh et al 1995; Silversand et al 1996; Bell et al 1997) The physiological and structural roles of the LC-PUFA in the reproduction of fish are reasonably well documented (Fernandez-Palacios et al 2011) However, due to the complex interactions and the fact that these roles vary with the reproductive stage, that is, different in gonad development, spawning and fertilization (fecundity), embryonic development (egg quality, hatching) and larval development (yolk sac use/retention, survival), the picture is still far from complete It has been recognized for some time that a (severe) deficiency of n-3 LC-PUFA in the broodstock diet impairs reproduction in fish (Watanabe et al 1984a; Chou et al 1993; Fern andez-Palacios et al 1995; Almansa et al 1999) More recently the importance of ARA was also recognized (Bell & Sargent 2003) Reproductive success is influenced by not only the levels but also by the ratios between LCPUFAs in the broodstock diet, the gonad and gametes (Fern andez-Palacios et al 2011) However, little is known about the physiological role of EPA and DHA during gonad development, ovulation and fertilization For ARA, it is known that ARA-derived eicosanoids, in particular the series prostaglandins (PGE2 and PGF2a), are important in the control of oocyte maturation and ovulation (Sorbera et al 2001; Kagawa et al 2003), are probably involved in embryogenesis (Bruce et al 1999) and larval development (Izquierdo & Koven 2011) and play a role in spermiation (Asturiano et al 2000) Kagawa et al (2003) showed in A japonica oocytes in vitro that PGF2a enhanced DHPinduced ovulation Indomethacin, actinomycin D and cycloheximide blocked DHP-induced ovulation and PGF2a reversed the effects of these inhibitors Similar effects of the series PGs were observed by Sorbera et al (2001) with in vitro Dicentrarchus labrax oocytes These authors further showed that addition of free ARA induced maturation of the oocytes Free ARA also enhanced GTH-induced maturation, while free EPA and DHA had the opposite effect ARA is also known to stimulate testicular testosterone in Carassius auratus testis in vitro through its conversion to PGE2 Again both EPA and DHA blocked the steroidogenic action of ARA and PGE2 (Wade et al., 1994, cited by Fern andez-Palacios et al 2011) The timing of spermiation may be delayed causing reduced fertilization rates due to depressed steroidogenesis caused by broodstock EFA deficiency or imbalance (Izquierdo et al 2001) In larval fish during endogenous feeding, ARA is selectively retained and has been shown to enhance survival and stress response (Tandler et al 1995) Fuiman & Ojanguren (2011) found in Sciaenops ocellatus no relation with egg FA profile and larval survival and growth, but did show a strong positive effect of the egg ARA level on the predator avoidance (escape) behaviour of the larvae Based on the work on Pagrus major (Watanabe et al 1984b), Sparus aurata (Fern andez-Palacios et al 1995; Tandler et al 1995) and Paralichthys olivaceus (Furuita et al 2000), the minimal amount of n-3 LC-PUFA in the broodstock diet seems to be 15-20 g kgÀ1 diet, with a minimal level of DHA no more than 6–7 g kgÀ1 diet and a minimum DHA/EPA ratio of 0.6 Lower levels of total n-3 LC-PUFA reduced fecundity, fertilization, hatching and Aquaculture Nutrition 19; 1–24 ª 2013 John Wiley & Sons Ltd survival Both Fernandez-Palacios et al (1995) and Furuita et al (2002) found that also higher levels of n-3 LC-PUFA (>20 g kgÀ1) impaired reproduction However, in these studies, both EPA and DHA were increased, EPA even more than DHA in the study by Fernandez-Palacios et al (1995) In studies with Dicentrarchus labrax (Bruce et al 1999) and Hippoglossus hippoglossus (Mazorra et al 2003), it was shown that DHA levels up to 40 g kgÀ1 diet did not impair reproduction Based on this, it seems that the optimal range for EPA is much narrower, from 7–10 to c 15 g kgÀ1 diet Consequently, the optimal DHA/EPA ratio will vary from 0.6 to 3, depending on the DHA level In A japonica, although there was no notable effect of FA composition of broodstock diet on the DHA content of the eggs, low-quality eggs contained significantly less DHA in the PL (Fig 3) (Furuita et al 2006; Ozaki et al 2008) There are indications that the optimal EPA range is furthermore also dependent on the dietary ARA level, and vice versa (Fernandez-Palacios et al 2011) Most fish-oilbased broodstock diets have ARA levels of 0.5–1.2 g kgÀ1 diet, with EPA/ARA ratios of 8–15 Positive effects, in particular on ovulation (fecundity) and fertilization, have been observed in Dicentrarchus labrax (Bell et al 1997; Bruce et al 1999), Hippoglossus hippoglossus (Mazorra et al 2003) and Gadus morhua (Salze et al 2005; Norberg et al 2009; Sawanboonchun 2009) with an increase in dietary ARA (up to 2–3 g kgÀ1 diet, with a concomitant decrease in EPA/ARA ratio till 1.5–6) Already in the early work of Watanabe et al (1984a,c) with Pagrus major, it could be noted that their cuttlefish meal diet, and the eggs from these fish, contained higher levels of ARA than their fishmeal-based diet and eggs Improved fertilization, hatching and larval survival were also observed in Paralichthys olivaceus with increasing ARA levels from to g kgÀ1 in Figure Effect of DHA level in polar lipid (PL) on egg quality (Furuita et al 2006) Lines are for hatching: y = 3.57 9À43.15 (P < 0.01), and for survival to days posthatch: y = 2.43 À39.25 (P < 0.05) NB The arrows for survival and hatching seem to have been exchanged in the original publication (Reprinted with permission of John Wiley and Sons) Aquaculture Nutrition 19; 1–24 ª 2013 John Wiley & Sons Ltd the broodstock diet (Furuita et al 2003c) A further increase in ARA dietary levels to 12 g kgÀ1 diet gave the opposite effect Similarly, negative effects of higher ARA levels in both the PL and the NL of A japonica eggs have been shown (Furuita et al 2003a, 2006) (Fig 4), although this is a bit puzzling in view of the fact that this ARA was largely synthesized by the eels themselves These data indicate that while ARA is essential for larval development, excess ARA levels can be detrimental for embryonic and larval development Broodstock origin and diet did not have a large effect on egg ARA level of A japonica (Furuita et al 2003a) (Fig 4a), although Ozaki et al (2008) did find lower levels of ARA in eggs of farmed A japonica fed with fish-oil-based diets One notable exception to the above is formed by krill, which is often fed, either frozen or as meal incorporated in the diet, to broodstock fish with excellent results (Watanabe et al 1985b; Watanabe & Kiron 1995; Mazorra et al 2003) Not only has krill a low level of ARA (c g kgÀ1 or 0.8% of FA), but also a very high level of EPA (c 25 g kgÀ1 or 20% of FA), resulting in DHA/EPA and EPA/ARA ratios of 0.5 and 25, respectively Also Furuita et al (2006) reported some positive effects of frozen krill in A japonica broodstock, but egg FA composition was not affected Apart from the levels and ratios of DHA, EPA and ARA, ‘pollution’ by other FA in farmed fish has been suggested as a cause for impaired egg quality Almansa et al (1999) reported for Sparus aurata that a high 18:1/n-3 HUFA ratio in both the NL and PL of the eggs negatively affected the fertilization rate Such an effect does not seem likely in eels, as even the eggs from wild eels show a high 18:1/n-3HUFA ratio (Table 2) Also negative effects of high levels of 18:2 in the eggs, as reported by Bell et al (1997) and Palacios et al (2011), not seem to occur in eels (Fig 5) Protein and amino acids Nitrogen-containing nutrients, protein and amino acids, nucleic acids, and more, form a large part of fish eggs (Table 1), but contrary to liposoluble compounds (and minerals, section Vitamins), there is no specific storage for these nitrogen-containing nutrients in the body There are a few studies on the effect of N-containing nutrients in broodstock nutrition in fish species who keep feeding during (part of the) gonad development It has been shown that defatted squid or cuttlefish meal has a positive effect on reproduction of Pagrus major (Watanabe et al 1991b) and Sparus aurata (Tandler et al 1995; Fern andez-Palacios et al 1997), but whether (a) (b) Figure Effect of ARA in PL on fertilization of A japonica eggs (a) (Furuita et al 2003a) and of ARA in NL on fertilization, hatching and survival (b) (Furuita et al 2006) Lines in (b) are for fertilization: y = À18.38 +75 (P < 0.05), for hatching: y = À30.31 +54.16 (P < 0.01) and for survival to days posthatch: y = À19.35 +35.63 (P < 0.05) NB in (b): AA = ARA (Reprinted with permission of Springer Science & Business Media and John Wiley and Sons) te hi oy Fe Bu la vu O ng d 20 n tc 10 40 tio 15 60 Ha 20 80 t 25 za Bonito oil-fed group ili 30 (b) 100 an Sunflower oil-fed group rt 35 Percentage (%) % total fatty acids (a) Figure Fatty acid profile in the total lipids of eggs (a) and reproductive performance (b) of A japonica fed with different oils Data from Yamada et al (2006) this was caused by the protein and/or the mineral fraction is not known In Colisa lalia, Shim et al (1990) found that deletion of certain amino acids from the broodstock diet reduced spawning performance and hatchability Fish fed the methionine deficient diet completely failed to spawn Also Harel et al (1995) showed that Sparus aurata broodstock fed a wheat gluten based diet (low in lysine) had significantly lower VTG levels, resulting in a decrease in larval survival by 50% In Plecoglossus altivelis, additional tryptophan in the broodstock diet advanced spermiation in males and final maturation in females, while a serotonine depletor delayed gonad development (Akiyama et al 1996) Matsunari et al (2006) showed for Seriola quinqueradiata broodstock that addition of 10 g kgÀ1 of taurine to the diet during months before spawning improved the spawning performance, expressed as the percentage of females spawning, from nil to 86% Taurine content of the ovaries was not different between treatments In A japonica, Higuchi et al (2012) showed that although taurine is essential in spermatogenesis, it is of endogenous origin, through DHP-stimulated biosynthesis from cysteine in the testes Gonzalez-Vecino et al (2004) reported that broodstock diets enriched with nucleotides improved the first feeding success and survival of Melanogrammus aeglefinus larvae Whether this was through incorporation in the eggs or through enhanced parental physiology was not reported The bulk (85–90%) of the (total) amino acids in the ovary and the eggs comes from VTG, the remainder largely from the choriogenins, or zona pellucida (ZP) precursor proteins (Pati~ no & Sullivan 2002; Lubzens et al 2010) The amino acid composition of VTG (and therefore largely of the eggs) is highly conserved in teleost fish (Fig 7a), although the most abundant amino acid, alanine, seems to be even more dominant in anguillid VTG (Hara et al 1980; Komatsu et al 1996) Compared with VTG, the ZP proteins of A japonica are relatively low in lysine and high in proline and glycine (Sano et al 2010) Aquaculture Nutrition 19; 1–24 ª 2013 John Wiley & Sons Ltd grouper, Epinephelus aeneus (Saint-Hilaire); rainbow trout, Oncorhynchus mykiss (Walbaum); Atlantic cod, Gadus morhua (L.); barramundi, Lates calcarifer (Bloch) and yellowtail kingfish, Seriola quinqueradiata (Temminck & Schlegel) (Azevedo et al 1998, 2005; Rodehutscord & Pfeffer 1999; Watanabe et al 2000a,b; Lupatsch et al 2003; Bureau et al 2006; Hatlen et al 2007; Glencross 2008; Booth et al 2010; Helland et al 2010) Many authors have demonstrated the influence of salinity on growth performance of fish (Gaumet et al 1995; Woo & Kelly 1995; Dutil et al 1997) Previous studies considering the effect of salinity on fish growth and metabolism have yielded controversial results Higher growth and lower metabolic rates were recorded at isotonic salinities (Gutt 1985; Febry & Lutz 1987; Waller 1992; Watanabe et al 1993; Lambert et al 1994), and this phenomenon was hypothesized to be caused by a reduction in the metabolic cost of iono- and osmoregulation In contrast, results from Morgan & Iwama (1991) and Sampaio et al (2001) showed that rearing fish at an isotonic salinity did not enhance growth or reduced metabolic rates Moreover, Tang et al (2006) reported that there are no differences in metabolic rate of juvenile turbot at different salinities after 48 h of acclimatization Turbot is considered to be a relatively euryhaline species and can adapt to salinities ranging between 10 and 40 g LÀ1 with a minimum level reported to be at 5–6 g LÀ1 (Waller 1992; Gaumet et al 1995; Person Le-Ruyet 2002; Tang et al 2006) Furthermore, it is well known that many juveniles prefer intermediary salinities, as found, for example, in estuaries, tidal coastal areas or coastal lagoons (Gaumet et al 1995; Boeuf & Payan 2001) Results from Imsland et al (2001) also suggested enhanced growth of juvenile turbot at an intermediate salinity (20 g LÀ1), especially at the thermal range of 18–22 °C Although several studies measured oxygen consumption and growth performance in turbot at different environmental salinities (Waller 1992; Gaumet et al 1995; Tang et al 2006), no data are available on how these changes affect the maintenance energy requirement as well as the efficiency of energy utilization for growth In the current study, the oxygen consumption (OC) (metabolic rate; including the energy expenditure for spontaneous activity) of fed fish was studied, because it is considered to be the metabolic state most closely representative of land-based aquaculture (MacIsaac et al 1997) This is also termed ‘routine oxygen consumption’ (ROC) (Jobling 1994) The present study intended to determine DEm, MEm and kg (DE), kg (ME) of juvenile turbot based on OC and using increasing feeding levels In addition, the effect of various salinities (10, 20, 30 g LÀ1) on OC as well as DEm, MEm and kg (DE), kg (ME) was studied Juvenile turbot, descendent from a domestic Norwegian broodstock, were used for the experiment and received from the hatchery Maximus A/S (Bedsted, Denmark) Fish were transferred to the rearing facilities of Gesellschaft f€ ur Marine Aquakultur, GMA (B€ usum, Germany) on 28th September in 2010 A fish meal–based diet was formulated to be similar to a commercial turbot feed considering gross energy (GE), macro- and micronutrients Referring to Kaushik (1998a) and Peres & Oliva-Teles (2008), the diet was formulated to meet the indispensable amino acid (IAA) requirements of turbot Titanium dioxide (TiO2) was added to the diet (10 g kgÀ1) as an inert marker to determine the digestibility of nutrients and energy Detailed information about the formulation and chemical composition of the experimental diet is given in Table The diet was supplied as pellets (4 mm diameter; pellet press 14-175; AMANDUS KAHL GmbH & Co KG, Hamburg, Germany) The experiment was carried out in 10 cylindrical tanks (250 L, bottom surface: ~0.52 m2) of a flow-through respirometer system (Fig 1) at Gesellschaft f€ ur Marine Aquakultur, GMA (B€ usum, Germany) Tanks were continuously supplied with water from a recirculation system equipped with sedimentation tanks, trickling biofilter, sump (~1500 L) and pump, whereby each tank had a separate inand outflow Flow rate through the tanks was adjusted to be 150 L hÀ1 For each tank, in- and outflow could be individually regulated by valves controlled by a central computer system, thus, a single record of oxygen concentration, temperature and flow rate was possible for every tank The same sensor unit was used for every measurement Additionally, each tank equipped with an extra circulation pump [Ocean Runner (OR) 2500; AB Aqua Medic GmbH, Bissendorf, Germany] with separate in- and outflow to sustain a constant mixing of the water body and to cluster uneaten feed or Aquaculture Nutrition 19; 135–150 ª 2013 John Wiley & Sons Ltd Table Diet formulation and chemical composition of the experimental diet (g or MJ kgÀ1) Diet formulation Herring fish meal LT1 Fish oil1 Wheat starch (pregelatinized)2 Wheat gluten2 Vitamin premix3 Mineral premix4 Titanium dioxide (TiO2)5 Chemical composition Dry matter (DM) In DM Crude protein (N 6.25) Crude lipid Crude ash Gross energy (GE) Ca P TiO26 748 75 80 80 3.5 3.5 10 933 586 164 160 21.8 41 23 10 Vereinigte Fischmehlwerke Cuxhaven GmbH & Co KG, Cuxhaven, Germany; [fish meal made from herring offal: DM, Crude protein, crude lipid, crude ash = 948; 724, 127, 114 g kgÀ1, respectively; LT, low-temperature cooking and drying (80–90 °C)] € ner Sta €rke – H Kro € ner GmbH, Ibbenbu € ren, Germany; Kro (Wheat gluten: DM, Crude protein = 920, 780 g kgÀ1, respectively) MicroMineral (supply per kg of diet): 1.4 mg CoSO4; 8.75 mg MnSO4; 1.75 mg CaI2; 8.75 mg CuSO4; 0.09 mg Se Vitamin standard (supply per kg of diet): 3500 I.U vitamin A; 700 I.U vitamin D3; 140 mg vitamin E (alpha-Tocopherol-acetate); 14 mg vitamin K3 (Menadione); 14 mg vitamin B1; 28 mg vitamin B2; 14 mg vitamin B6; 0.03 mg vitamin B12; 210 mg vitamin C (Monophosphate); 28 mg vitamin B5; 140 mg vitamin B3; 5.60 mg Folic Acid; 0.35 mg Biotin; 140 mg Inositol; 700 mg Betain-anhydrate; 98 mg ZnSO4; 140 mg Etoxyquin (contents of vitamins, minerals and additives as specified by the manufacturer) Kronos Titan GmbH & Co.oHG, Nordenham, Germany Calculated by: TiO2 = Ti 1.67 faeces to the central settling funnel at the bottom of the tanks Uneaten feed and faeces could be removed by an outlet valve To minimize external influences inducing stress, both sides and the back of the tanks are made of opaque PVC The front is made of transparent PVC allowing observation of the fish Detailed information and descriptions of the respirometer system were given by Stiller (2010) A restricted ration level (RRL) experiment realized at three different salinities (10, 20, 30 g LÀ1) including three feeding levels (I–III) was designed Feeding levels I–III corresponded to a daily feed supply of 0.3, 0.6, 0.9 g kgÀ1 of metabolic body weight (MBW) Nine days prior the start of the trial, 78 juvenile turbot were devided into nine balanced Aquaculture Nutrition 19; 135–150 ª 2013 John Wiley & Sons Ltd experimental groups (six fish each) having a similar BW The remaining fish as well as the experimental groups were stocked in separate 50-L aquaria at a salinity of 10 g LÀ1 and fed the experimental diet at feeding level III once daily for acclimatization to experimental conditions Before transferred to the nine respirometer tanks, turbot of the experimental groups were individually weighed again to ensure similar BW of groups Mean initial BW in each respirometer tank was 985 Æ 10 g (164 Æ g BW per fish) After transferring to the respirometer, fish were acclimated to the tanks at a salinity of 10 g LÀ1 for days Experimental groups were fed once a day per hand (11:30–12:15 hours) with three replications at every feeding level To ensure that all supplied feed was consumed the central collecting funnel at the tank bottom was covered during feeding Prior every feeding faeces were eliminated from the tanks Additionally, all uneaten pellets were removed from the respirometer tanks nearly 10 after feeding and quantified To calculate the actual feed intake, the average weight of one dry pellet (determined prior the start of experiment) was multiplied with the number of uneaten pellets Fish not used for the respirometry trial were still stocked in the aquaria at similar salinities like the experimental groups and fed at feeding level III After acclimatization of the experimental groups, ROC was measured at the salinity level of 10 g LÀ1 for a 24-h period Thereafter, fish were starved for days, and OC [starving oxygen consumption (SOC)] was recorded (24 h) The same procedure was repeated at salinities of 20 as well as 30 g LÀ1, hence resulting in three successive measuring cycles Every cycle was followed by days of adaption to the next higher salinity Therefore, feeding continued and salinity was increased stepwise within days to left over days for acclimatization at the target salinity prior the next measuring cycle The lower and intermediate salinities (10 and 20 g LÀ1) were obtained by mixing tap water with seawater (North Sea, B€ usum, Germany) and the higher salinity (30 g LÀ1) by adding commercial seasalt for fish keeping (‘Preis sea salt’; Preis-Aquaristik, Bayerfeld, Germany) to seawater During the experiment, temperature was on average 16.5 Æ 0.2 °C and dissolved oxygen ranged between 8.9 and 6.6 mg LÀ1 Thus, the oxygen concentration never deceeded the minimum oxygen (6.0–7.0 mg LÀ1) required for maximum growth of turbot (Boeuf et al 1999; Person Le-Ruyet 2002) The mean total ammonia nitrogen (TAN) concentration was 0.1 Æ 0.1 mg LÀ1 and never exceeded 0.2 mg LÀ1 Mean nitrite (NO2) was 0.4 Æ 0.3 mg LÀ1 with a maximum of 1.0 mg LÀ1 The photoperiod was set on a 14 h L:10 h D cycle (daybreak at 06:00 hours) The total experiment lasted 35 days Figure General design of the respirometer system (© Stiller 2010) Recirculation pump, Manometer, Water cycle, Overpressure valve, Fresh water inflow, Respirometer (fish tank), Overflow tube, Sedimentation barrel, Sedimentation tank, 10 Sump, 11 Trickling filter, 12 Sampling tube to the sensors, 13 Merger of sampling tubes, 14 Sensors (pH, CO2, O2), 15 Online control unit, 16 Main switch, 17 Online control, 18 Data transfer Fish were weighed individually at the beginning and at the end of the experiment to calculate the average MBW per fish at the OC measurements based on the specific growth rate Fish starved at least day prior to weighing To determine OC, the oxygen concentrations in the respirometer tanks as well as flow rate and temperature were automatically recorded by the computer system every hour in each experimental group for a 24-h period Thereby, the respirometer tanks were successively measured (time interval per tank: min) within hour and the time at the end of this cycle taken to be representative for each tank One tank was left without fish and used as reference representing the inflow oxygen concentration by the system for all experimental tanks within the hourly record cycle as well as to correct for OC by microorganism inhabiting the system To avoid any effect of fouling processes on oxygen concentration, the reference tank was carefully cleaned before the start of each 24-h measuring period Prior to every 24-h measurement, the sensor (JUMO dTRANS O2 01; JUMO GmbH & Co KG, Fulda, Germany) was calibrated on humidified air (saturation 100%; immediately above the water surface) To ensure an adequate water quality as well as to verify the recording of the computer system temperature (OxyGuardHandy Polaris 2; OxyGuard International A/S, Birkerød, Denmark), salinity (HI 96822 digital seawater refractometer; HANNA instruments Inc., Woonsocket, RI, USA), TAN and NO2 (colorimetric test kit – Microquant; Merck KgaA, Darmstadt, Germany) were manually measured every day Following the respirometer trial, the experimental fish and the remaining turbot from aquaria were used for determination of nutrient and energy digestibility of the diet Fish were divided in three subsets Subsets consisted of two replicates and were acclimated to salinities of 10 (n = 12 fish per replicate), 20 (n = 12 fish per replicate) and 30 g LÀ1 (n = 15 fish per replicate), respectively At each salinity level, fish were fed 12 days once daily with 1.0 g kgÀ1 of MBW Approximately 12 h after the last feeding, the content of the posterior intestine (from the ileocaecal valve to the anus) was carefully sampled by dissection once fish were killed by an overdose of MS222 (200 mg LÀ1) No abrasion of intestinal walls was performed to minimize the contamination of digesta with epithelial cells Due to limited quantities to accurate chemical analyses, the sampled intestinal contents were pooled for each salinity level, stored in a plastic container at À18 °C and subsequently freeze-dried Prior to chemical analyses, intestinal samples were homogenized According to Naumann & Bassler (1976), the experimental diet and intestinal contents [due to the small quantity except for crude lipid (CL)] were analysed in duplicate for dry matter (DM), crude ash (CA), CL and GE (adiabatic bomb calorimeter C200; IKA-Werke GmbH & Co.KG, Staufen, Germany) CL analysis was carried out with HCl hydrolysis Crude protein (CP, N 6.25) was determined using the Dumas combustion principle (True Aquaculture Nutrition 19; 135–150 ª 2013 John Wiley & Sons Ltd Specâ N; LECO Corporation, St Joseph, MI, USA) TiO2 in the diet and intestinal contents was determined according to Brandt & Allam (1987): TiO2 was solubilized by the Kjeldahl procedure for h in 0.96 g gÀ1 sulfuric acid, then 0.35 g gÀ1 hydrogen peroxide (H2O2) was added to the filtered TiO2 solution to form a yellow complex and colour intensity measured in a spectrophotometer (visible spectrophotometer 6300; Barloworld Scientific Ltd T/As Jenway, Dunmow, Essex, UK) at 405 nm The determination of calcium and phosphorus contents of the diet was performed according to VDLUFA VII 2.2.2.6 method (VDLUFA 2008) Specific growth rate (SGR), % per day: 100 (ln BW1Àln BW0) daysexpÀ1, where daysexp are the experimental days, and BW0 and BW1 are initial and final mean BW, respectively Metabolic body weight, kg0.8: BWt0.8, where BWt represent the average BW (kg) of fish at each 24-h measuring period BWt was calculated based on SGR According to our previous statements (Dietz et al 2012), the common metabolic weight exponent for energy of 0.8 was applied Apparent digestibility coefficients (ADC), %: 100 [1À (% feedTiO2/% faecesTiO2) (% faecesn/% feedn)], where subscript n stands for a specific nutrient or energy To calculate energy intake from digestible protein (DEIp), the energetic equivalent of 23.6 kJ gÀ1 protein (Blaxter 1989) was used Non-faecal N excretion was calculated using the ratio between retained protein (RP) and digestible protein intake (DPI) from a previous experiment using turbot descendent from the same broodstock and a similar diet (Dietz et al 2012) The non-faecal N excretion was derived from the difference between digestible N intake (DPI/6.25) and N deposition (RP/6.25) The energy loss associated with this excretion was calculated using the energetic equivalents of 24.9 kJ gÀ1 ammonia N and 23.1 kJ gÀ1 urea N (Elliott & Davison 1975) Assuming 15% of N is excreted as urea, and 85% as ammonia (Dosdat et al 1995) results in 24.6 kJ gÀ1 non-faecal N Metabolizable energy intake (MEI, kJ kgÀ0.8 dayÀ1) was defined as: Digestible energy intake (DEI, kJ kgÀ0.8 dayÀ1) – energy loss by non-faecal N (EN-loss, kJ kgÀ0.8 dayÀ1) Oxygen consumption (mg hÀ1) was calculated according to the following equation: OC ¼ ðcOr2 ÀcOi2 Þ Â F In this equation, cOr2 and cOi2 are the concentrations of dissolved oxygen (mg LÀ1) in the reference tank (corresponding to Aquaculture Nutrition 19; 135–150 ª 2013 John Wiley & Sons Ltd inlet concentration by the system) and individual respirometer tanks (corresponding to outlet concentrations), respectively F defines the flow rate (L hÀ1) through the individual respirometer tanks ROC and SOC reflecting the average OC per MBW for fed and starving fish within 24 h (mean of 24 measurements per tank per salinity level), respectively To consider the unsteady-state situation in the flow-through system, the calculated OC-values were verified by applying the following equation proposed by Niimi (1978): Qo2 ¼½ððFÂc0 ÂeÀF=VÂt Þ À ðFÂcÞÞ=ð1 À eÀF=VÂt ފ þ ðF  cin Þ (1) where, referring to the present study, QO2 represent OC, c0 and c are cOi2 of two consecutive records, F = 150 L hÀ1, V = 250 L, t is the time intervall between the records of c0 and c (=1 h) and cin the mean between cOr2 related to records of c0 and c, respectively Specific dynamic action (SDA, mg kgÀ0.8 hÀ1): SDA = ROCÀSOC SDA represents the OC above starving and reflects the energy requirement accociated with feeding, for example, searching, ingestion, digestion and absorbtion of feed Heat (energy used for SDA and maintenance inclusive spontaneous activity) was calculated based on OC (13.6 kJ per g O2 consumed; Elliott & Davison 1975) All data considering energy metabolism presented per MBW Digestible energy intake, MEI and RE (RE = MEIÀHeat) at feeding levels I–III were subjected to linear regression analysis for determination of DEm, MEm, kg (DE) and kg (ME) according to the following linear regression equations: RE ðkJ kgÀ0:8 dayÀ1 Þ ¼ a þ kg ðDEÞ Â DEI ðkJ kgÀ0:8 dayÀ1 Þ RE ðkJ kgÀ0:8 dayÀ1 Þ ¼ a þ kg ðMEÞ Â MEI ðkJ kgÀ0:8 dayÀ1 Þ Energy intakes at RE = are defined as the maintenance requirements, and the efficiencies of energy utilization for growth above maintenance are given by the slopes However, this statistical approach does not provide the standard errors for the estimates of the energy maintenance requirements, which are necessary for testing the significance of differences Therefore, DEm and MEm were alternatively determined by the following equations, where energy intake is taken as dependent and RE as independent variable: DEI ðkJ kgÀ0:8 dayÀ1 Þ ¼ DEm þ 1=kgðDEÞ Â RE ðkJ kgÀ0:8 dayÀ1 Þ À0:8 À1 MEI ðkJ kg day Þ ¼ MEm þ 1=kgðMEÞ Â RE ðkJ kgÀ0:8 dayÀ1 Þ Here, DEm and MEm are represented by the y-intercepts, and the slopes 1/kg (DE) and 1/kg (ME) – the reciprocals of the kg -values – describe the costs for DE or ME (kJ) per unit of RE (kJ), respectively Because DEm, MEm, kg (DE) and kg (ME) based on ROC, it must be noted that DEm and MEm -values determined by extrapolation at RE = include also the energetic requirements for spontaneous activity All the data were checked for homogeneity of variance using box plots and residual plots DEI, MEI and RE were subjected to analysis of covariance (ANCOVA), and multiple comparisons were used to compare slopes (Holm) and y-intercepts (Tukey) of the linear regression equations, respectively Thereby, the influence of tank as random factor could not be included in the statistical model Results of growth and energy budget parameters were analysed by ANOVA applying linear mixed models (y ~ salinity + feeding level + salinity feeding level + tank as random factor) regarding heterogenous variances and followed by all-pair comparisons between salinities within feeding levels as well as between feeding levels within salinities Differences were considered to be significant at the P < 0.05 level All statistical data analyses were performed using R-Software version 2.11.1 (R Development Core Team 2010) Results are presented as mean Æ SD Table Apparent digestibility coefficients (ADC)1 (%) of nutrients and energy in juvenile turbot at different salinities Salinity (g LÀ1) 10 20 30 Dry matter Organic matter2 Crude protein Crude ash Energy 73.4 83.7 86.6 19.4 87.1 69.8 80.5 84.2 13.5 84.3 68.9 79.8 83.8 11.6 83.9 ADC: calculated using one pooled sample of intestinal content of fish fed at salinities of 10 (n = 24), 20 (n = 24) and 30 (n = 30) g LÀ1 Calculated in diet and intestinal contents by: Organic matter = dry matter – crude ash The deviations between results of ROC and SOC determined using the actual data of OC and their estimates based on the unsteady-state mass balance (Eq 1) are [...]... anguilla, L.) females during induced sexual maturation In: Aquaculture Europe 11 18– 21 October 2011 Rhodes, Greece Aquaculture Nutrition 19; 1–24 ª 2013 John Wiley & Sons Ltd McEvoy, L., Holland, D & McEvoy, J (1993) Effect of spawning fast on lipid and fatty acid composition of eggs of captive turbot (Scophthalmus maximus L.) Aquaculture, 114, 131–139 McKenzie, D.J., Higgs, D.A., Dosanjh,... biosynthesis underlies a tradeoff between reproduction and flight capability in a wing-polymorphic cricket Proc Natl Acad Sci U.S.A., 99, 16829–16834 Aquaculture Nutrition 19; 1–24 ª 2013 John Wiley & Sons Ltd Aquaculture Nutrition 2013 19; 25–38 doi: 10.1111/anu.12086 1,2,3 1 1,2 3,4 3,4 1,2 CIIMAR/CIMAR – Centro Interdisciplinar de Investigacßa~o Marinha e Ambiental, University... an increase in all dietary indispensable amino acids influenced the secondary stress response in handled Senegalese sole by minimizing Aquaculture Nutrition 19; 25–38 ª 2013 John Wiley & Sons Ltd Aquaculture Nutrition 19; 25–38 ª 2013 John Wiley & Sons Ltd 18.8 Æ 2.3 1.1 Æ 0.3 89.2 Æ 1.4 13.7 Æ 2.7 10.8 Æ 1.4 1.3 Æ 0.4 91.0 Æ 4.6 11.6 Æ 3.5 9.0 Æ 4.6 60 days 15.7 Æ 4.0 15 days... composition of ovaries from wild fish and ovaries and eggs from captive fish of white sea bream (Diplodus sargus) Aquaculture, 216, 299–313 Celino, F.T., Yamaguchi, S., Miura, C & Miura, T (2009) Arsenic inhibits in vitro spermatogenesis and induces germ cell Aquaculture Nutrition 19; 1–24 ª 2013 John Wiley & Sons Ltd apoptosis in Japanese eel (Anguilla japonica) Reproduction, 138, 279–287 Cerda,... Takeuchi, T (2002) Effects of high levels of n-3 HUFA in broodstock diet on egg quality and egg fatty acid composition of Japanese flounder, Paralichthys olivaceus Aquaculture, 210, 323–333 Aquaculture Nutrition 19; 1–24 ª 2013 John Wiley & Sons Ltd Furuita, H., Ohta, H., Unuma, T., Tanaka, H., Kagawa, H., Suzuki, N & Yamamoto, T (2003a) Biochemical composition of eggs in relation to egg... A-induced bone deformity model Aquaculture, 315, 26–33 Hamre, K (2011) Metabolism, interactions, requirements and functions of vitamin E in fish Aquacult Nutr., 17, 98–115 Hamre, K., Waagbø, R., Berge, R.K & Lie, Ø (1997) Vitamins C and E interact in juvenile Atlantic salmon (Salmo salar, L.) Free Radical Biol Med., 22, 137–149 Aquaculture Nutrition 19; 1–24 ª 2013 John Wiley & Sons Ltd Hamre,... effectively incorporated in the ovaries For As, Boyle et al (2008) found a decrease Aquaculture Nutrition 19; 1–24 ª 2013 John Wiley & Sons Ltd in egg production and hatchability of Danio rerio fed naturally contaminated Nereis diversicolor Total As levels can be quite high in fishmeals and fish oils used in aquaculture feeds (Sloth et al 2005) These authors claimed that this is of little concern,... in devel- Aquaculture Nutrition 19; 1–24 ª 2013 John Wiley & Sons Ltd oping eggs and yolk-sac larvae of walleye pollock Theragra chalcogramma Fish Sci., 72, 620–630 Ohkubo, N., Sawaguchi, S., Nomura, K., Tanaka, H & Matsubara, T (2008) Utilization of free amino acids, yolk protein and lipids in developing eggs and yolk-sac larvae of Japanese eel Anguilla japonica Aquaculture, 282, 130–137... Physiol., Part A Mol Integr Physiol., 119, 787–793 Rønnestad, I., Koven, W., Tandler, A., Harel, M & Fyhn, H.J (1998b) Utilisation of yolk fuels in developing eggs and larvae of European sea bass (Dicentrarchus labrax) Aquaculture, 162, 157–170 Rønnestad, I., Thorsen, A & Finn, R.N (1999) Fish larval nutrition: a review of recent advances in the roles of amino acids Aquaculture, 177, 201–216 Salze, G.,... Bass (Dicentrarchus labrax) Biol Reprod., 64, 382–389 Aquaculture Nutrition 19; 1–24 ª 2013 John Wiley & Sons Ltd Southgate, P.C., Lee, P.S & Rimmer, M.A (1994) Growth and biochemical composition of cultured Sea Bass (Lates calcarifer) Larvae Asian Fish Sci., 7, 241–247 Støttrup, J.G., Jacobsen, C., Tomkiewicz, J & Jarlbæk, H (2013) Modification of essential fatty acid composition in broodstock