1. Introduction
Among the alternatives to fish meal (FM) as protein sources, fish by-product meal is considered safe and has good nutritional value [
1], and different types have been used in the diets of several species of fish. For example, shark by-product meal has been utilized in red sea bream,
Pagrus major [
2], and in yellowtail,
Seriola quinqueradiata [
3]; salmon by-product meal in red sea bream [
4]; and tuna by-product meal (TBM) in spotted rose snapper,
Lutjanus guttatus [
5], Korean rockfish,
Sebastes schlegeli [
6], olive flounder,
Paralichthys olivaceus [
7], and greater amberjack,
Seriola dumerili [
8]. Although TBM offers a sustainable high-protein ingredient for aquafeeds, it can influence the palatability of formulated diets, potentially affecting intake and overall performance [
8]. Indeed, lower feed intake has been reported in some species due to this issue, although the specific component in TBM that affects palatability has yet to be identified [
7,
8]. To address this challenge, the incorporation of feeding stimulants has emerged as a promising strategy to improve the attractiveness and consumption of feed.
Roe, which is the term used to describe fish eggs collected in skeins [
9], is not only a rich source of protein and unsaturated fatty acids but also contains free glutamate and nucleotide acids [
10,
11]. The existence of these components leads to a strong umami intensity in roe-based products. Glutamic acid stimulated positive feeding behavior in crustaceans [
12], and the effective stimulation of the appetite due to nucleotides such as inosine monophosphate (IMP) was reported in other aquatic animals [
13]. In Japan, salted (shiozuke) and soy sauce-marinated salmon roe is highly popular. During the processing of this product, the roe sacs are gently massaged in warm salted water to separate the individual eggs from the membrane, which are then thoroughly rinsed in cold water to remove impurities and excess membrane residue. However, many salmon roes are damaged and dissolved in the water during production. This water is assumed to contain both glutamic acid and nucleotides. Therefore, powdered salmon roe processing liquid (PSRL) possibly contains umami components and may have high value. In this study, PSRL was selected as a feed attractant based on a preliminary study (unpublished) on red sea bream, which showed a positive effect on feed intake. Feed attractants ultimately improve growth by increasing feed palatability, feed intake, and digestion and absorption [
14]. However, the potential of PSRL as a feeding stimulant in diets where FM is partially replaced by TBM remains largely unexplored.
The greater amberjack (
Seriola dumerili), which is widely farmed in the Mediterranean region and Japan, grows relatively quickly [
15] and has a high market value. An earlier study on this species suggested that only 14.5% (equivalent to 25% when a post hoc test alone was used) of FM protein could be replaced by TBM due to the lower food intake resulting from the palatability issue [
8]. A major reduction in growth performance was found in previous research on the same species when 50% of FM protein was replaced by TBM [
8]. Therefore, the replacement levels in the current study were set at 25% and 40%. It was hypothesized that a feeding stimulant might be able to increase the acceptance of TBM-based diets, which could help to raise the FM replacement level. Therefore, in this study, we aimed to further determine effective degrees of replacement of FM with TBM and to evaluate the impact of PSRL supplementation on the acceptance of diets containing TBM, providing information on its applicability in aquafeed formulations.
2. Materials and Methods
2.1. Products and Nutritional Composition
The PSRL used in this study was provided by RegenWorks Co., Ltd. (Fukushima, Japan). It was produced via drying roe processing water, which contained damaged roe liquid. The composition of the TBM and its processing methods have been discussed elsewhere [
8]. In brief, after removing some of the bones, raw materials were dried and pulverized to prepare TBM powder. Comparisons of proximate composition and amino acid (AA), free AA, and fatty acid (FA) contents among FM, TBM, and PSRL are provided in
Table 1,
Table 2,
Table 3 and
Table 4 respectively. The lowest protein content was found in PSRL, whereas TBM exhibited the highest protein and lipid contents. Phosphorus (P) and calcium levels were higher in FM (
Table 1). The total content of indispensable amino acids (IAAs) and dispensable amino acids (DAAs) was lower in PSRL compared to the other ingredients (
Table 2). Similarly, the total free IAA content was lower in PSRL; however, it contained higher levels of glutamic acid, which is known as an umami component (
Table 3). In PSRL, the total saturated FA (SFA) content was lower, whereas the monounsaturated FA (MUFA) content was higher compared to that in the other ingredients. However, there was no major variation in total polyunsaturated FA (PUFA) content between PSRL and FM (
Table 4).
2.2. Dietary Formula and Composition
A total of seven experimental diets, including a control diet (C), were prepared. The formulae and proximate compositions are shown in
Table 5. In a previous study [
8], growth performance significantly declined when more than 25% of FM-derived protein was replaced with TBM protein. Based on these findings, FM protein replacement levels were set at 25% and 40%, with PSRL included at two different degrees. The diets in which FM-derived protein from the C diet was replaced at 25% and PSRL was added at 0%, 5%, and 10% were referred to as T25R0, T25R5, and T25R10, respectively. Similarly, diets in which FM-derived protein was replaced at 40% and PSRL was added at 0%, 5%, and 10% were designated as T40R0, T40R5, and T40R10, respectively. After thoroughly mixing all ingredients with water, the diets were pelletized using a machine (12VR-750SDX, Alpha Royal Co., Ltd., Osaka, Japan) and dried at 60 °C for 24 h. Crude protein and crude lipid contents ranged from 48.2% to 49.7% and 13.1% to 14.4%, respectively.
Table 6 presents the AA composition of the test diets. There were no major differences in IAA or DAA content among them. The levels of all free IAAs and DAAs, except for histidine and proline, were higher in the test diets than in the C diet (
Table 7).
As the amount of TBM increased, the contents of SFAs and MUFAs tended to increase, while the content of PUFAs tended to decrease (
Table 8).
2.3. Fish Husbandry and Sampling
Seven hundred greater amberjack juveniles were purchased and stocked in a 3000 L circular indoor rearing tank for acclimation. During this two-week period, the juveniles were fed a commercial diet containing 52% crude protein (Marubeni Nisshin Feed Co., Ltd., Tokyo, Japan). For distribution into experimental tanks, the fish were anesthetized using 250 ppm phenoxyethanol (Wako Pure Chemical Industries Ltd., Osaka, Japan) after a 24 h fasting period. The amberjack juveniles, with an average weight of 1.74 g, were allocated to 21 experimental tanks (30 juveniles per 500 L tank), with three replicates per treatment. To determine the initial whole-body proximate composition, a sample of 30 juveniles was stored at −80 °C. They were fed twice daily (08:30 and 14:30) until apparent satiation, six days per week for six weeks. Filtered seawater was supplied to the experimental tanks at approximately 5 L/min. Water quality (dissolved oxygen, temperature) was monitored every day at 10:30 and 16:00 using a DO meter (Pro2030, YSI Nanotech Co., Ltd., Kawasaki, Japan). The water temperature and dissolved oxygen concentration were 26.7 ± 1.2 °C (mean ± SD) and 5.6 ± 0.7 mg/L, respectively, throughout the experiment. The photoperiod was set to 12 h of light (07:00–19:00) and 12 h of darkness (19:00–07:00). During the experimental period, the tank bottoms were cleaned daily at 11:00 using a siphon. Any dead fish were immediately removed and weighed.
During the rearing experiment, fish from all tanks were weighed in the pool at two-week intervals after being anesthetized with 250 ppm phenoxyethanol (Wako Pure Chemical Industries, Ltd., Osaka, Japan). On the final day of the experiment, three fish from each tank were dissected to measure the relative weights of the whole internal organs, liver, stomach, and pyloric caeca. Additionally, five fish from each tank were sampled to analyze whole-body proximate composition. Blood was collected from three fish per tank using a heparinized syringe and centrifuged at 3000 g for 15 min at 4 °C, and the separated plasma was stored at −80 °C until analysis.
The following formulae were used to calculate growth parameters, relative organ weight, and retention efficiency:
Weight gain, WG (%) = 100 × (final mean weight − initial mean weight)/initial mean weight;
Specific growth rate, SGR (%/day) = 100 × (ln final weight − ln initial weight)/rearing period (days);
Daily feeding rate, DFR (g/100 g fish/day) = 100 × total feed intake/[(mean of initial and final number of fish × mean of initial and final body weight) × rearing period];
Feed efficiency, FE (%) = 100 × [total wet weight gain (g)/total dry feed intake (g)];
Condition factor, CF = 1000 × (W/L3), where W = wet body weight (g) and L = fork length (cm);
Survival rate (%) = 100 × final number of fish/initial number of fish in a tank;
Relative weight (%) of viscera (VSI), liver (HSI), stomach (SSI), pyloric caeca (PSI), and intestine (ISI) = 100 × [wet weight of viscera, liver, stomach, pyloric caeca and intestine (g)/wet body weight (g)];
Retention efficiency (%) of protein (PRE) and lipids (LRE) = 100 × [(final whole-body protein and lipid − initial whole-body protein and lipid)/total protein and lipid intake].
2.4. Biochemical Analyses
AOAC methods [
16] were used to analyze the proximate composition of ingredients, diets, and whole-body fish samples. The AA and free AA contents and FA composition in the ingredients and diets were analyzed by the Japan Food Research Laboratory (Osaka, Japan). In brief, the fatty acid (FA) content of the experimental diets and fish whole bodies was determined according to Folch et al. [
17] using a gas chromatograph (GC4000, GL Science, Tokyo, Japan) equipped with a capillary column (InterCap-Pure-WAX, GL Science). The amino acid (AA) content of the diets was determined using high-performance liquid chromatography (HPLC; GL7700, GL Science, Tokyo, Japan) according to the method described by Teshima et al. [
18]. Plasma chemical components, including total protein (TP), glutamic oxaloacetic transaminase (GOT), and glutamic pyruvic transaminase (GPT) activities, glucose (GLU), total cholesterol (TC), and triglyceride (TG), were measured using commercial kits (Fuji Dry-Chem, Fujifilm Company Ltd., Tokyo, Japan).
2.5. Statistical Analysis
Data are expressed as the mean ± standard deviation. Growth parameters were based on triplicate samples, while relative organ weights and plasma parameters were measured using nine samples per treatment. The effects of TBM and PSRL levels, as well as their interactions among treatments, were determined using two-way ANOVA. Significant differences among dietary treatments, including the C diet, were analyzed using one-way ANOVA followed by Tukey’s test (p < 0.05). Statistical analyses were conducted using the SPSS program for Windows (v. 10.0).
3. Results
A comparison of growth results among treatments at the end of the experiment is shown in
Table 9. The FMW, WG, and SGR followed the same pattern across treatments, with no significant differences in these parameters among experimental groups C, T25R0, and T25R5 (
p > 0.05). However, the other experimental groups showed significantly lower values than the C group (
p < 0.05). Similarly, there were no significant differences in DFR among groups C, T25R0, T25R5, and T25R10, whereas the other experimental groups had significantly lower values compared to the C group (
p < 0.05). However, no significant differences were observed in survival rate or FE among treatments (
p > 0.05). Results from the two-way ANOVA analysis indicated that the FM replacement level with TBM had a significant effect on FMW, WG, SGR, and DFR, but there was no significant effect on survival rate and FE. Similarly, PSRL and its interaction with TBM did not have a significant effect on either of these growth parameters.
The CF and relative organ weight values for each experimental group are presented in
Table 10. No significant differences in these factors were observed among the treatments (
p > 0.05). Moreover, neither TBM, PSRL, nor their interaction had a significant effect on relative organ weight compared to the C diet (
p > 0.05).
Table 11 shows the initial and final whole-body proximate compositions and P contents. Although all test diets exhibited significantly lower crude lipid content compared to the C group (
p < 0.05), no significant differences were found in the other parameters among the treatments (
p > 0.05). The two-way ANOVA indicated a significant effect of replacing FM with TBM (
p < 0.05); however, neither PSRL nor its interaction with TBM showed a significant effect on those parameters compared to the C diet (
p > 0.05).
No significant difference in PRE was observed among the treatments (
Table 12,
p > 0.05). However, fish fed diets T25R0, T40R0, T40R5, and T40R10 exhibited significantly lower LRE compared to those fed the C diet (
p < 0.05). Although TBM had a significant effect on LRE, neither PSRL nor its interaction with TBM significantly affected PRE or LRE.
Table 13 presents the plasma constituents in greater amberjack juveniles at the end of the rearing trial. No significant differences among the treatments were observed, nor were there any significant effects of TBM, PSRL, or their interaction on the plasma parameters (
p > 0.05).
4. Discussion
As mentioned earlier, in a previous study on greater amberjack juveniles when the optimal replacement level was determined using polynomial regression analysis, it was found that only 14.5% of FM protein could be replaced by TBM [
8]. However, when the same data were analyzed using the Tukey post hoc test—the same method used in the present study—the replacement level was found to be 25% [
8]. It was assumed that the poor growth observed in fish fed diets with more than 25% FM replacement in the previous study was due to a reduced DFR caused by palatability issues. Therefore, in this study, we aimed to enhance the palatability and DFR of a TBM-based diet by adding a feed attractant, PSRL. Unfortunately, this failed to stimulate the DFR in greater amberjack juveniles, regardless of supplementation dose. The observed trend of this rate was consistent with the findings of the previous study on the same species [
8], reinforcing the conclusion that only 25% of FM protein can be replaced by TBM. The results of the two-way ANOVA indicate that only the level of TBM in the diet significantly affected growth, while PSRL supplementation and its interaction with TBM had no significant effect.
Although AA imbalance is often considered a contributing factor to poor DFR and growth [
19], dietary AA was not responsible for the low growth observed in greater amberjack juveniles when similar TBM was used in the previous study [
8]. In the present research, all IAA levels in the TBM-based diets were similar to or higher than those in the control diet. There is very limited information on AA requirements in greater amberjack. From the available information, lysine levels in TBM-based diets ranged from 3.51% to 3.66%, exceeding the reported need for 2.03% to 2.11% for this species [
20]. In addition, all free AAs, except histidine, were more abundant in TBM-based diets compared to those in the control diet. From the available information, it has been suggested that the essential FA requirements for this species can be met when n-3 PUFAs constitute 12 g/kg of the diet [
21]. In this study, the levels of n-3 PUFAs in the TBM-based diets ranged from 21.9% to 25.7%, equivalent to 27.9 g/kg to 32.8 g/kg of the diet (FA content in diet = total dietary lipid × 0.892) [
22], thereby exceeding the reported requirement. Moreover, these levels were higher than the recommended levels of 12–17% of total FA in the feeding stage of
Artemia [
23] and 1.0–1.7 g/100 g diet in the broodstock stage [
24] in this species. Therefore, both AA and FA levels in the TBM-based diets were unlikely to be responsible for the poor growth observed in greater amberjack juveniles.
Therefore, it appears that a lower DFR, which may be due to the palatability problem, is the main reason for the poor growth observed with TBM-based diets, as diet plays a crucial role in obtaining nutrients. Feed attractants are known to address palatability issues, enhance the DFR by increasing appetite, and ultimately stimulate growth [
14]. Various feed attractants have shown positive effects in different aquatic species. For example, squid paste has been effective in inducing feeding in Chinese perch,
Siniperca chuatsi [
25], Chinese soft-shelled turtle,
Pelodiscus sinensis [
26], and black tiger shrimp,
Penaeus monodon [
27]; squid paste, glutamic acid, and dimethyl-beta-propiothetin have improved appetite, growth, digestion, and absorption in Chinese mitten crab,
Eriocheir sinensis [
28]; and krill meal, krill oil, and fish hydrolysate have enhanced DFR in Pacific white shrimp,
Litopenaeus vannamei [
29]. Feed attractants not only increase DFR but also improve nutrient retention efficiency [
28]. However, in this study, PSRL did not demonstrate a feed-attracting effect or significantly enhance nutrient retention efficiency.
Glutamic acid and nucleotides are generally considered as umami components that influence feeding behavior in fish. The strong umami intensity of roe from marine organisms is primarily attributed to free glutamate and nucleotide acids [
10,
11,
30]. Glutamic acid has been reported to influence feeding behavior and feed utilization [
12,
28,
31], while nucleotides such as guanosine monophosphate (GMP) and inosine monophosphate (IMP) are known to effectively stimulate appetite [
13,
28]. The so-called equivalent umami concentration (EUC) is used to quantify the umami intensity in an ingredient or food. Yamaguchi et al. [
32] proposed an empirical equation for the EUC based on the concentrations of free glutamate and nucleotides such as IMP, adenosine monophosphate (AMP), and GMP. According to this equation, in a review by Mouritsen [
33], the EUC values of various raw and processed marine products are summarized. For instance, Alaska pollock roe, bonito flakes, and sea urchin have EUC values of 3500 [
34], 28,000 [
35], and 60,000 [
36], respectively, while salmon roe shows a much lower EUC of only 500 [
37]. Although water-soluble extracts prepared from salmon roe by Hayashi et al. [
37] contained free IMP, AMP, and GMP, resulting in an EUC of 500, these nucleotides were not detected (detection limit: 0.01%) in the PSRL used in the current study. Therefore, the EUC value of PSRL used in this research could not be calculated. Despite higher levels of free glutamic acid in PSRL compared to those in FM or TBM, the absence of free IMP, AMP, and GMP suggests that the EUC of PSRL is likely even lower than that of salmon roe reported by Hayashi et al. [
37]. Variation in EUC values may be attributed to differences in raw materials, processing methods, egg maturity stages, fish maturity, diets used to feed the broodstock, species, and the stability of PSRL during feed processing. For example, glutamic acid content has been reported to increase with the maturity of salmon roe [
38], and the chemical properties of roe can vary even within the same female [
39]. Further studies are necessary to clarify whether nucleic acids play an important role as a feed attractant in this species.
In this research, the final whole-body lipid content and LRE were significantly affected by TBM inclusion in the diet. However, in a previous study on the same species, no adverse effects were observed in whole-body lipid content or LRE with up to 75% FM replacement by TBM [
8]. Additionally, no significant differences in protein or lipid content were found with up to 100% FM replacement using the same TBM product, suggesting that lipid digestibility was not the cause of reduced LRE or whole-body lipid content in the present research. One notable difference is the smaller size of fish (1.7 g) used in the current study compared to that in the previous one (6.7 g). Therefore, the variation in lipid deposition and LRE may stem from differences in the energy utilization of lipids in fish of different sizes or from reduced assimilation due to a lower DFR. This hypothesis is supported by the absence of significant differences in plasma TC levels across treatments. Bile acids, synthesized from TC, play a key role in lipid digestion and absorption [
40]. If lipid digestion is impaired, bile acids are expected to be more mobilized, which may affect the TC level in plasma. Thus, the stable plasma TC levels suggest that digestibility was not a limiting factor for lipid deposition in this study. No significant changes were observed in other plasma parameters typically used to assess potential adverse effects of FM replacement. Plasma TP is often an indicator of health and immune status, while GOT and GPT levels are markers of liver function [
41]. Neither TBM inclusion nor PSRL supplementation affected these parameters, suggesting that the health status of the fish was not compromised. Similarly, in a previous study, no changes in plasma constituents were observed, except for those in TG levels, with up to 50% FM replacement by TBM in greater amberjack [
8].