Hybridization between fishes is a practical and effective tool for improving the production of fish; the technique has also been used with domestic animals since prehistoric times. Hybrid striped bass Morone chrysops × M. saxatilis, tilapia Oreochromis spp., and channel catfish Ictalurus punctatus × I. furcantus are economically available, attributable to the rapid growth of industrial aquaculture production worldwide (Swann et al., 1994; Brecka et al., 1995; Riche et al., 2004; Rawles et al., 2006, Phelps et al., 2007; Ponzoni et al., 2007).

Red sea bream Pagrus major (RSB) and black sea bream Acanthopagrus schlegeli (BSB) are major aquaculture species in East Asia, China, Japan, and Korea. However, these two porgies are classified into different genera, and differ in their ecology and physiology, such as habitats, avoidance of ultraviolet radiation, development of sensory organs, energy allocation patterns, intestinal microflora, and protandrous sex reversal (Muroga et al., 1987; Kinoshita and Tanaka, 1990; Chang et al., 1995; Mana and Kawamura, 2002; Tang et al., 2003; Fukunishi et al., 2006; An et al., 2008). Harada (1991) and Murata (1998) found that the hybrid (F₁) between female RSB and male BSB had higher tolerance against lower water temperature, dissolved oxygen (DO), and salinity than RSB. Murata (1998) investigated the genetic and cytogenetic characteristics of F₁ and found the same number of chromosomes between F₁ and parent species, and that F₁ had hybrid sterility. Additionally, Murata (1998) and Kim et al. (2011) showed faster growth of F₁ juveniles at one year old, compared with RSB and BSB. It seems likely that the F₁ will be an appropriate aquaculture species in Korea and East China where the seawater temperature often falls below 10℃, resulting in mass mortality of RSB. However, there are few reported studies regarding the nutritional requirement of F₁ for establishing an aquaculture industry (Kim et al., 2009a, 2009b).

Dietary protein and lipid levels are essential factors for efficient fish production (National Research Council, 1993). It is known that some hybrids have similar or different dietary protein and lipid requirements compared with their parents (Brecka et al., 1995; Twibell et al., 2003). Moreover, there are some suggestions that the hybrids showed heterosis in growth, survival and disease tolerance, distinguishable from their parent species (Logan, 1968; Blanc and Chevassus, 1982; Tuncer et al., 1990; Brecka et al., 1995; Moreau and Pauly, 1999; Harel and Place, 2003).

Given this background, this investigation was conducted to determine a suitable dietary protein/lipid ratio for early F₁ juveniles, which would be useful information for further promoting and developing the F₁ aquaculture industry in East China and Korea.

The F₁ were produced artificially by cross-breeding of mature female RSB of the Kinki University strain (4 years old, 4.5 ± 0.4 kg, n = 5) and male BSB (2 years old, 310 ± 89.4 g, n = 4), reared in the Fish Nursery Center of Kinki University, Uragami and Shirahama, in May 2006. The eggs and milt from RSB, of the same group used for F₁, were artificially fertilized to produce RSB. After hatching, F₁ and RSB larvae were fed with rotifer Brachionus plicatilis and Artemia salina (INVE Aquaculture, Dendermonde, Belgium) from 3 to 29 and from 22 to 40 days after hatching (dAH), respectively. The rotifer and Artemia were enriched with n-3 highly unsaturated fatty acids (n-3 HUFA), using a commercial enrichment product (Marine Glos; Nisshin Marinetech Co., Yokohama, Japan). The juveniles from 28 dAH were fed with commercial pellets (Marubeni Nisshin Feed Inc., Tokyo, Japan), increasing pellet diameter with their growth, until the start of the trials.

Five diets were prepared with various ratios of crude protein/lipid (CP/CL)—60/7, 55/12, 51/17, 46/23, and 41/28—as shown in Table 1. Brown fish meal (crud protein and lipid contents were ca. 69.2 and 9.4%, respectively) and wheat gluten and fish oil were supplied as the main protein and lipid sources, respectively. All experimental diets were prepared by mixing the ingredients appropriately and adding 30% tap water externally, and were then pelletized using a laboratory pellet machine. The pellets were freeze-dried and sieved to obtain two particle sizes for the two separate trials using F₁ and RSB of different sizes. The pellet particle size was 355-425 µm for smaller F₁ and RSB juveniles, weighing 0.32 ± 0.01 g and 0.26 ± 0.01 g at 43 and 41 dAH, respectively. For larger F₁ and RSB juveniles, weighing 3.67±0.06 and 4.03±0.07 g at 75 and 73 dAH, respectively, pellets of 1.9 mm in diameter were prepared and freeze-dried. The diets for both trials were prepared just before the start of each trial and stored in a freezer at ‒20℃ until used.

There were two trials in this study. Both trials were conducted under a photoperiod cycle of 14-h light and 10-h dark.

In Trial 1, 50 smaller juveniles of F₁ and RSB each were introduced into 40-L rectangular tanks. This trial was conducted in triplicate for each diet, and juveniles were fed four times daily until apparent satiation, at 08:00, 11:00, 14:00, and 17:00-h, for 2 weeks. Water temperature and DO were 20.6 ± 0.7℃ and 6.5 ± 2.1 mg/L, respectively.

In Trial 2, 20 larger juveniles of F₁ and RSB each were introduced into 40-L rectangular tanks. This trial was also conducted in triplicate for each diet, and juveniles were fed three times daily, at 09:00, 13:00, and 17:00-h, until apparent satiation for 4 weeks. Water temperature and DO were 24.3 ± 2.2℃ and 7.6 ± 1.7 mg/L, respectively.

At the beginning and end of each trial, juveniles were weighed and sampled after fasting for 24-h. Growth performances of the fishes were evaluated at the end of each trial and compared among the treatments. Growth performance was estimated using daily feeding rate, survival, specific growth rate (SGR), feed efficiency (FE), condition factor (CF), protein efficiency ratio (PER), and apparent protein, lipid, and energy retention efficiencies (PRE, LRE, and ERE). To evaluate the whole body proximate composition, 10 juveniles from Trial 1 and 5 from Trial 2 were collected from each tank. Also, six juveniles from each tank were sacrificed and dissected for various organ samples to measure hepatosomatic indices (HSI) and viscerosomatic indices (VSI). All samples were immediately transferred to a freezer (‒40℃) until analysis.

At the end of Trial 2, blood samples were obtained from the caudal aorta of three fishes per tank with heparinized syringes. Hematocrit levels and hemoglobin content (Hb) were determined using a microhematocrit method (Brown, 1980) and a commercial kit (Wako Pure Chemical Industries, Osaka, Japan), respectively. For assay of hepatopancreatic glutamateoxalacetate transaminase (GOT) and glutamate-pyruvate transaminase (GPT) activities, the hepatopancreas removed from juveniles was homogenized with nine volumes of deionized water (v/w) using a Potter-Elvehjem glass homogenizer. After centrifugation of the homogenate (10,000 g, 15 min), the resulting supernatant was used for assaying GOT and GPT activities using a commercial kit (Wako Pure Chemicals Industries). Enzyme protein was assayed according to Lowry et al. (1951). The enzyme assay was conducted at 30℃ for 20 min, and the activity was expressed as international units per mg protein, as specific activity.

The proximate compositions of the diets and whole bodies were assayed using Association of Official Analytical Chemists methods (1995). Sugar contents of diets were measured by the phenol-sulfuric acid method (Hodge and Hofreiter, 1962). The energy concentrations of diets and whole bodies were determined directly using an automatic oxygen bomb calorimeter (IKA-Werke, Staufen, Germany).

Data obtained from both trials were subjected to a two-way analysis of variance (ANOVA) to evaluate the effect of experimental diets, fish species, and interactions. Where significant differences were found, the means within and among the treatments were further compared using Duncan’s new multiple range test followed by one-way ANOVA at the P < 0.05 significance level (Harter, 1960). All statistical analyses were conducted using the SPSS software (ver. 17.0 for Windows; SPSS Inc., Chicago, IL, USA).

The growth performance data from Trial 1 are presented in Table 2. Two-way ANOVA showed that different dietary CP/ CL ratios had significant effects on final mean body weight, daily feeding rate, weight gain, SGR, and FE (two-way ANOVA, P < 0.05). All growth parameters except SGR showed significant differences between F₁ and RSB, where F₁ showed significantly higher performances in final mean body weight, weight gain, FE, and PER compared with RSB (two-way ANOVA, P < 0.05).

In contrast, final mean body weight, weight gain, SGR, and FE showed decreasing trends with decreasing dietary CP/CL ratios for both species (P < 0.05). For F₁, the final mean body weight, weight gain, SGR, and FE on 60/7 and 55/12 diets were significantly higher than those on 46/23 and 41/28 diets, while all growth parameters on the 60/7 diet were significantly higher than those on 51/17, 46/23, and 41/28 diets for RSB (P < 0.05). Overall, the survival rate was significantly higher in F₁ than RSB (P < 0.05).

There was no difference in carcass proximate compositions between F₁ and RSB at the start of the rearing trial (Table 3). At the end of Trial 1, two-way ANOVA showed significant differences in all proximate composition parameters, while both diet and species and their interaction had significant effects on apparent nutrients and energy retention (P < 0.05). Generally, nutrient retention values were higher in F₁ than RSB, LRE on the 60/7 diet.

Although there was no difference in whole body protein content or PRE among the dietary treatments in F₁ (P > 0.05), whole body lipid and energy contents exhibited a significantly increasing trend with decreasing dietary CP/CL ratio, with lower LRE and ERE in both F₁ and RSB (P < 0.05). The PRE, LRE, and ERE of RSB decreased markedly with decreasing dietary CP/CL ratio, being significantly higher on the 60/7 diet than the 46/23 or 41/28 diets (P < 0.05).

At the end of Trial 2, the experimental diets had significant effects on final mean body weight, weight gain, SGR, FE, and PER for both species (two-way ANOVA, P < 0.05) (Table 4). Also, survival, SGR, weight gain, FE, and PER showed significant differences between F₁ and RSB (two-way ANOVA, P < 0.05), where F₁ showed higher trends in almost all cases than RSB.

There was no significant difference in final mean body weight, daily feeding rate, survival, weight gain, SGR, or FE among the dietary treatments in F₁. In contrast to Trial 1, the PER of F₁ showed an increasing trend with decreasing dietary CP/CL. The final mean body weight, weight gain, and SGR of RSB were significantly higher on the 60/7 diet than the other diets (P < 0.05). However, the PER increased with decreasing dietary CP/CL, as did F₁ (Table 4).

The final whole body proximate compositions, except ash content, were significantly affected by dietary treatments in both species (two-way ANOVA, P < 0.05; Table 5). However, the whole body protein, lipid and energy contents were not significant difference between the fish species (two-way ANOVA, P > 0.05). Both diet and fish species had significant effects on LER and ERE, with higher values in F₁ than RSB except ERE on the 60/7 and 55/12 diets (two-way ANOVA, P < 0.05).

With decreasing dietary CP/CL, whole body protein showed decreasing trends, but lipid and energy showed increasing trends in F₁ and RSB. LRE also tended to decrease with decreasing CP/CL in both species. While F₁ fed 60/7, 55/12, and 51/17 diets had significantly higher LRE than the 41/28 diet (P < 0.05), RSB fed the 60/7 diet showed significantly higher LRE than with the other CP/CL diets (P < 0.05). Significantly higher ERE was observed with 55/12, 51/17, 46/23, and 41/28 diets for F₁, and the 55/12 diet for RSB (P < 0.05) (Table 5).

Regarding hematological parameters, only Hb contents were affected by fish species, and RSB tended to be lower than F₁ (two-way ANOVA, P < 0.05) (Table 6). The CF, HSI, and VSI were affected significantly by diet and by fish species (two-way ANOVA, P < 0.05), where the indices tended to lower in F₁ than RSB. Additionally, HSI and VSI showed a tendency to increase with decreasing dietary CP/CL in both species. Although hepatopancreatic GOT activity was not affected by dietary treatment or fish species, hepatopancreatic GPT activity was significantly lower in F₁ (two-way ANOVA, P < 0.05).

The establishment of mass juvenile production for finfish aquaculture industry requires various technical disciplines and basic knowledge related to the biological, chemical, physiological, and environmental aspects of each aquaculture fish. In particular, understanding the relationship between changes in nutritional requirements and biochemical metabolism in the early life stages of fish will produce rapid improvement and success in mass fingerling production. Thus, this study was designed to identify a suitable dietary protein and lipid content and/or ratio for early F₁ juveniles, which are expected to be a promising new aquaculture species in Korea and eastern China where the water temperature becomes very low in the winter.

It is well-known that the survival of fingerling production depends on whether the dietary needs for early juvenile stages are met (National Research Council, 1993; Takeuchi, 2001). In both rearing trials, survival rates were significantly different between species, with markedly higher values in F₁ than RSB (P < 0.05) (Tables 2 and 4). These results may be because F₁ had greater adaptable and performance under these experimental conditions (e.g., diet formulation, rearing density) than RSB.

There are few reports dealing with suitable dietary protein and lipid levels in early juvenile stages of F₁ or RSB. We were obligated to end the rearing Trials 1 and 2 at 2 and 4 weeks after the start due to the juveniles’ fast growth, and the resulting over-density in the rearing tanks. However, the final mean body weights in each trial increased by 2- to 3-fold relative to the initial mean body weight within this rearing period. Thus, it is reasonable to conclude that the 2- to 4-weeks rearing period for the early juvenile stages of these species are not too short to obtain reliable data.

The growth performance data revealed that suitable CP/ CL proportions for F₁ and RSB were 55/12 and above 60/7, respectively, in Trial 1, whereas they were lower, at 41/28 for F₁ and above 51/17 for RSB, in Trial 2. These results for RSB are consistent with the results of other authors (Yone et al., 1974; Takeuchi et al., 1991). Takeuchi et al. (1991) and Yone et al. (1974) reported that RSB weighing 1.6-29.0 g required 52-55% dietary protein and 15% dietary lipid. Watanabe et al. (1984) showed that mature RSB, weighing 689 g, required 45% dietary protein, which is markedly lower than juveniles (Takeuchi et al., 1991). In contrast, Arakawa et al. (1980) indicated a suitable dietary protein level of 40% for BSB weighing 2.9 g, which has been reported to be lower than RSB. A two-way ANOVA revealed significant differences in growth performance and nutrient retention between the two fish species. Generally, F₁ performed better on diets with low CP/CL (46/23, 41/28) than did RSB, providing evidence that F₁ had a lower protein requirement than the female parental fish (RSB). These findings are consistent with those in several other species where hybrid fish ordinarily showed lower protein requirements than the maternal fish (Millikin, 1982, 1983; Berger and Halver, 1987; Shiau and Huang, 1989; Tuncer et al., 1990; Brown et al., 1992; Swann et al., 1994; Brecka et al., 1995; Twibell and Brown, 1998).

Azevedo et al. (2004) found a significant difference in growth between rainbow trout and Atlantic salmon by feeding various CP/CL diets. Regarding the cause of the difference, they considered that it was due to the difference in nitrogen and energy retention efficiency between the two species. Differences in capabilities between fish species were also identified in this experiment, and effects in growth improvement by substituting fish meal for fish oil were observed in both juvenile F₁ and RSB. However, significant differences (P < 0.05 by two-way ANOVA) between the two fish species were found in PRE, LRE, and ERE (Tables 3 and 5). In particular, the retention efficiency of F₁ in the low CP/CL diets (46/23, 41/28) showed a markedly higher tendency than RSB. These results suggest that F₁ experiences a protein-sparing effect when fed diets with higher lipid levels, as reported for porgy (Vergara et al., 1996; Schuchardt et al., 2008). However, the juvenile F₁ has a greater capacity to synthesize and accumulate protein using dietary lipid as an energy source than does RSB.

From the hepatic transaminase activity, used as an indicator of the breakdown and consumption of amino acids, the results showed a reduced tendency in GPT activity in both species with a reduced dietary CP/CL ratio. However, there was a significant difference between F₁ and RSB, with lower activity in F₁ than RSB. It has been reported that feeding a high-lipid diet to fish decreases levels of glycogen, blood nitrogenous compounds, gluconeogenesis, and glycolysis enzymes (e.g., PFK, G6Pase, G6P DH) activities, and GPT activity, along with an increase in hepatic fat content. Thus, gluconeogenesis and glycolysis, fatty acid synthesis, and amino acid decomposition are inhibited by the ingestion of greater quantities of lipid (Shimeno, 1983). This is the so-called “protein-sparing effect” of lipid that decreases nitrogen excretion by inhibiting the breakdown of amino acids and improves the accumulation of dietary protein in the body by using dietary lipid as an energy source (Shimeno et al., 1981; Skalli et al., 2004). Although the amino acid balance in these experimental diets was not considered, the lower GPT activity in F¹ may indicate that this fish has greater protein-sparing capacity than RSB through lipid mobilization. This may be attributable to the rapid growth of F₁, as has been reported previously (Murata, 1998; Kim et al., 2011). However, further studies are necessary to clarify this issue.

In conclusion, these results demonstrate that the hybrid F₁ is a promising aquaculture species because of its superior growth performance and higher survival rate even at a lower dietary CP/CL ratio. The results also indicated that mass juvenile production of F₁ will be more economical than that of RSB.