4. Discussion
The challenge of gradually replacing FM and FO in aquaculture feeds must be addressed using a holistic research strategy tailored to industry needs [
51]. This study aimed to improve knowledge on feed formulations that can be managed and well accepted by genetically selected GSB, and to assess the impact of different feeding strategies in long-term feeding up to commercial size.
The fish performance of GSB was not uniform, as FCR and growth were better with feeding CTRL diet, but SGR, FI, and K with the ALT diet. Different studies have shown that the full or partial replacement of FM or FO independently with alternative ingredients is feasible without negatively affecting fish growth performance, with better results observed when combining multiple alternative ingredients [
11,
17,
18,
52,
53,
54]. Carvalho et al. [
8] highlighted, in GSB diets, that replacing FO with a combination of microalgae oil (AO), poultry oil (PO), and vegetable oil (VO) with 15% FM resulted in growth and SGR efficiency similar to the control diet with FO but growth performance decline accompanying the reduced FM percentage.
Research on replacing FM and FO with new and innovative ingredients is still in its early stages. Although several meta-analyses have investigated the replacement of FM and FO with novel ingredients in aquafeeds, none have assessed the combined use of multiple alternative ingredients within a single dietary strategy [
12,
15,
55,
56]. The substitution of FM and FO in commercial diets with alternative ingredients such as IM or single-cell protein and a mixture of PO and AO significantly reduced growth and productive parameters like SGR in short-term GSB studies [
25,
28]. This reduction is presumably due to alternative diets containing less than 15% FM, substituted by a single protein source, different from a vegetable meal (VM). Meagre (
Argyrosomus regius) fed an alternative diet of insect meal (IM), microalgae biomass, and oil from tuna canning water showed lower FCR but similar growth performance to that of the control diet, although the FM replacement rate was low [
57]. It must be considered that using new protein sources to replace FM may worsen zootechnical performance due to suboptimal amino acid profiles; diets with ingredients like poultry meal have lower apparent digestibility for some amino acids compared to diets with less FM substitution, especially in long-term feeding studies when fish reach commercial size [
58,
59]. The better growth indicators outcomes for ALT diet could involve enhanced amino acid availability, improved FI, and increased feed palatability [
11,
58].
Based on the genotype results, although there were no differences in FCR, the HG group achieved its intended improvement objective, showing significantly better growth parameters than the REF genotype (approximately 9%). Previous studies with genetically selected GSB [
25,
28] or in other species testing selected strains such as European seabass (
Dicentrarchus labrax) [
58,
60,
61] and rainbow trout (
Oncorhynchus mykiss) [
62] have also demonstrated better growth and productive parameters in the selected genotypes, regardless of dietary treatment. The ALT diet results in the HG group suggest that genetic selection can take advantage of novel dietary formulations using emerging ingredients. Previous short-term feeding trials have demonstrated that growth-based genetic selection, when combined with diets based on plant ingredients [
24] or novel ingredients, such as poultry and microalgae oils [
25], can enhance growth performance [
58,
63]. This suggests that genetically selected fish are more capable of adapting their metabolism, which is characterized by increased intestinal plasticity [
24,
28,
58,
64]. Specifically, GSB selected for high growth and fed alternative diets have demonstrated increased intestinal length to enhance nutrient absorption [
24], a more resilient gut microbiota [
64], improved digestive enzyme activity, and higher apparent digestibility coefficients (ADC) [
28] in comparison with non-selected fish.
In commercial farms, feed intake at AS is not considered an effective practice [
29]. Working in tanks with few fish, even at high densities, the AS hand-feeding protocol contributed to the lack of significant differences in FCR. In accordance with the results of this study, feed restriction has been shown to reduce growth parameters [
31,
65] even when applied cyclically or based on body weight [
30,
66]. However, when assessing the involvement of genetic selection, the HG genotype makes the best use of feeding at apparent satiety, while the REF genotype is most penalized with the 65AS strategy.
The fish flesh composition appears to be strongly influenced by dietary ingredients, but other parameters such as feed ration and fish size, as in the present study, affect the proximate composition of the fillet [
67]. This feed ration effect has also been described in short-term experiences with GSB feed on novel raw materials and involves genetic selection, where fillet proximal composition was not modified [
25]. Feeding strategies affect the body composition of fish, especially lipids stored in the liver, viscera, and muscles, which are the main source of energy for maintenance during periods of starvation or feed restriction [
65]. The fish fed AS showed a mean increase of 18.75% in lipids and a concomitant 1.63% reduction in fillet moisture content compared to those fed 65AS, an inverse relationship previously reported [
23,
67]. In fact, the application of feed restriction strategies decreases the percentage of lipids present in fillets, while the moisture content decreases when ration supply is high [
66,
68]. Additionally, genotype may modulate these effects, suggesting that HG GSB have a greater ability to conserve lipids in fillets under dietary restriction. In European seabass, differing results have been found regarding lipid content and genetic selection [
57,
59].
Regarding the decline processes during ice storage, it was not affected by treatments, and the reduction in protein percentage detected at 4 dph was due to the concomitant effect of proteolytic enzyme activity [
37,
69]. A decrease in sarcoplasmic protein content during the dripping process after protein denaturation [
23,
70] and variations in the proportion of protein content lead to a higher percentage of lipids.
Diet composition often significantly influences the cellular conformation of the fish, particularly the fatty acid composition [
71], as in the present study for all the experimental factors studied. Thus, the higher MUFA content in the fish fed the CTRL diet was related to the palmitoleic acid (C16:1
n-7) present in vegetable and fish oils [
72]. The
n-3LC-PUFA in the fillets from fish fed the ALT diet showed the connection with the high percentage of DHA provided by AO [
73], while the higher EPA content present in FO was accumulated to a greater extent in fish fed the CTRL diet. Despite the higher EPA content in the fillet from the CTRL diet, the overall EPA + DHA content in the fillet remained significantly higher in the fillets of GSB fed with the ALT diet. This is particularly relevant considering the importance from a human nutrition perspective in enhancing the reduction in cardiovascular risk, inflammation, and neurodegenerative diseases [
74,
75,
76]. Thus, per 100 g of fillet, the total mg of EPA + DHA provided by the fish fed the ALT diet was 477 mg, almost 25% more than the 397 mg of the fillets of the fish fed the CTRL diet, although in both cases above 250 mg/day recommended by the European Food Safety Authority [
77]. Although cooking can lead to a decline of the
n-3/
n-6 ratio, since SAT are relatively heat-stable under typical cooking temperatures and LC-PUFA are more susceptible to degradation due to their higher degree of unsaturation, the use of mild cooking methods, such as steam oven processing, can help preserve the lipid quality of the fillet [
35].
Not only does the EPA + DHA content in fish have beneficial health effects, but studying the complete fatty acid profile will provide information on the potential cardiovascular risks derived from the atherogenic and thrombogenic effects of SAT and the protective role of PUFA and MUFA [
45]. Thus, the lower atherogenicity index (AI) value detected in the fish fed the ALT diet was due to its higher level of PUFAs, especially
n-3PUFA, along with a lower content of myristic acid (C14:0). This reflects diet composition, with a higher contribution of myristic acid from fish oil in the CTRL diet. For
n-3PUFA, although the content in both diets was similar, the higher presence of DHA in the ALT diet due to the inclusion of microalgae oil, along with its selective accumulation [
78], causes the fish fed this diet to achieve higher proportions of
n-3PUFA in the fillet. In the case of thrombogenicity index (TI), no differences have been detected between diets, as the higher PUFA values in fish fed the ALT diet have been counterbalanced by a higher content of stearic acid (C18:0), largely contributed to the diet by PO [
79]. The low values obtained for both AI and TI, comparable to those already described for farmed gilthead seabream fillets [
80,
81], support the reduction in cardiovascular disease risk [
82], even though no recommended value has been established by public health organizations [
83]. This is likely because the platelet aggregation properties of polar and neutral lipid fractions depend on the fish species and their origin [
84].
Fish fed the ALT diet showed improved n-3 fatty acid levels without increasing n-6 fatty acids, resulting in a significantly higher n-3/n-6 ratio, addressing a key challenge of replacing marine ingredients like FO with n-6-rich alternatives such as PO. The novel ingredients in the ALT diet not only avoid negative impacts on zootechnical performance but also enhance the nutritional value of the fillet for consumers.
The fact that the REF genotype had a higher concentration of
n-3LC-PUFA, while the HG genotype displayed a greater proportion of MUFA, was also noted in a long-term study in European seabass. In this fish species, the selected genotype exhibited lower levels of
n-3LC-PUFA [
60] or exhibited higher levels of DHA in the fillet compared to a commercial genotype [
58], showing that the selective breeding for high growth has both positive and negative impacts on fish fillet quality [
25]. In rainbow trout, selective breeding for improved growth has led to changes in the fatty acid composition of the fillet, including a reduction in beneficial unsaturated fatty acids [
85]. Moreover, genetic selection can alter the expression of crucial genes involved in PUFA biosynthesis, thereby decreasing their proportion in the fillet [
86].
Regarding the feeding strategy effect, the significantly higher percentage of
n-3LC-PUFA in the fillet at 85AS and 65AS compared to AS has also been described in other species, such as lean strains of Atlantic salmon that accumulate more
n-3LC-PUFA than fat strains [
87].
n-3LC-PUFA can be selectively retained using different feeding levels to meet the FA requirements and tissue membrane function [
68,
88,
89]. The C18:1
n-9 proportion found in the AS group was also described in rainbow trout, establishing that starvation or a reduction in feed ration has been related to a decrease in the MUFA ratio due to the association with the decrease in lipid percentage in the fillet [
90]. The interaction between genotype and feeding strategy revealed that the MUFA content, including C18:1
n-9, in the HG fillets remained consistent regardless of the feeding strategy. In contrast, the GSB REF group fed the 85AS and 65AS strategies showed lower MUFA content than the other groups. This connects with the selected phenotypes’ improved endurance to food restriction, which allows them to retain dietary nutrients more efficiently. Currently, dietary management and genetic selection are the primary tools used to control the muscle fat content in farmed fish [
68].
The reduction in the percentage of
n-3LC-PUFA fatty acids, accompanied by an increase in the percentage of SAT throughout the ice storage, indicates that unsaturated fatty acids are more prone to oxidation due to the greater number of double bonds, a critical issue regarding the fatty acid composition of fish fillet [
67]. This oxidative process leads to the formation of compounds such as hydroperoxides and aldehydes, which can compromise the skeletal muscle structure and contribute to changes in fillet texture and overall quality, as reflected in our results [
91]. In line with this, Alexi et al. [
92] reported similar alterations in GSB, linking them to enzymatic degradation and fatty acid breakdown [
23]. Previous studies have shown that unlike wild GSB, the fillet of farmed GSB exhibits minimal variation in fatty acid composition during the ice storage period, regardless of the diet provided. Notably, the
n-3/
n-6 ratio remains favorable throughout the shelf life, thereby preserving the cardiovascular health benefits associated with fish consumption [
23].
In addition to zootechnical and nutritional aspects, sensory characteristics, including texture as perceived by consumers, play a crucial role in the acceptance and market value of the product. Notably, the sensory analysis results indicated that the experimental diets did not significantly affect the long-term organoleptic properties. However, the diets led to variations in the fillet fatty acid profiles, altering the concentration of volatile aldehydes derived from
n-3 or
n-6PUFA [
34], and even within the
n-3PUFA category, between EPA and DHA [
93]. However, replacing FO with AO and increasing DHA content have not shown a clear influence on sensory attributes, as previously confirmed in both Atlantic salmon [
94] and GSB [
95]. Only in diets where the inclusion of AO promotes large differences between EPA and DHA accumulations can sensory differences be detected [
96]. About the effect of genotype, the higher perceived juiciness intensity in HG fish is likely related to a higher moisture content in the fillet [
97], which is associated with the different growth rates.
Interestingly, the feeding strategy was identified as the factor with the greatest impact on the sensory profile. Differences detected in global odor and flavor intensity, the shiny appearance of the fillet, fatness and juiciness textures, and oily odor and flavor describe variations in fillet fat content [
98] of more than 20% in the fish fed AS versus 65SA. While for the degree of fat perception in the mouth during chewing, the relationship with fillet fat seems clear, juiciness is also positively correlated and depends on muscle fat content [
80,
99] because fish fat gives a soft and succulent, i.e., juicy, mouthfeel. Odor and flavor, assessed in both overall intensity and oiliness, are determined by the volatile products of lipid oxidation [
97], with fattier fish more prone to develop these characteristics compared to leaner fish. The other effect of the feeding strategy on sensory perception was a metallic flavor associated with increased dietary restriction (65AS). Considering that lipid oxidation is essential to produce a metallic flavor [
100], the higher
n-3LC-PUFA content in the fillet of fish fed a restricted diet conditioned its higher susceptibility to oxidative rancidity, a phenomenon that may be attributable to the presence of metal ions, which catalyze the oxidation of these fatty acids [
100].
Regarding fillet texture properties, flesh hardness can be influenced by reducing the force required for compression when fish meal (FM) is decreased in the diet [
101], particularly when the final FM content is below 10% and replaced with vegetable meal (VM) [
35]. However, substituting both FM and VM with other animal protein sources, as proposed in the ALT diet, retains and even enhances fillet firmness compared to fillets from the fish fed diets containing only 15% FM. The texture of the fish fillet is dependent on the amino acid ratio of the diet [
102], necessitating the supply of amino acids such as hydroxyproline, which is considered conditionally essential in fish [
103]. This ensures the correct structure and strength of connective tissue [
104], which cannot be guaranteed by a vegetable meal-based diet with low or no hydroxyproline contribution [
105]. Replacing soy protein concentrate in the control (CTRL) diet with poultry by-product meal has been effective in improving fillet firmness because the hydroxyproline content provided by poultry by-products is 35–45 times higher than that provided by soy [
106].
Growth in fish involves recruitment and hypertrophy of muscle fibers and modifications in muscle cellularity that promote variations in meat texture [
107] because connective tissue is relatively more abundant in a muscle with high fiber density, and higher values of texture parameters will be achieved [
108]. The HG group, which exhibited significantly greater growth than REF over the same period and at the same age, would have experienced hypertrophy to a greater extent. Because of the lower fiber density, the fillet is less cohesive and requires less force to break it up. A moderate feed ration restriction (85AS) has increased the fillet resistance to deformation compared to AS, supported by a low-fat muscle content. In fact, a significant loss of hardness and thus softening of the flesh has been associated with an increase in muscle fat [
109,
110], which is negatively correlated with the maximum force to compression [
111,
112].
Proteolytic activity during ice storage induces structural changes in fish fillets, leading to softening and the loss of freshness and quality [
113]. The narrow window studied in the present research, up to 4 dph, did not detect significant differences in filamentous proteins during ice storage, indicating resistance to proteolysis. The increased immunoreactivity of actin and m-calpain observed at 4 dph, particularly pronounced in the REF genotype, may be attributed to the activation of calpain proteases during the early post-mortem storage period, which coincides with the initial stages of muscle degradation in fish [
114,
115]. Furthermore, calpain activity degrades actin filaments, releasing α-actinin from the myofibrils into the cytoplasm, resulting in the loss of Z-disc integrity [
38] and increasing actin detection by specific antibodies. The loss of Z-disc integrity contributes to muscle degradation and results in a reduction in fillet firmness, which may negatively influence consumer acceptance [
67]. The differential response of calpastatin to m-calpain and μ-calpain in the REF genotype could indicate the role of calpastatin as an endogenous calpain inhibitor [
37] and the same differential effect of the HG group on growth and resilience to feed restriction. In Nile tilapia (
Oreochromis niloticus), calpains and calpastatin exhibited inversely coordinated expressions in response to starvation [
115]. In contrast to our findings on GSB, in which no dietary or gene group differences were observed, the endoproteases µ-calpain, m-calpain, and calpastatin remained unaltered throughout ice storage for up to 10 days [
37]. Only desmin showed degradation, but this occurred after more than 4 days of ice storage. Similar results were observed in seabass, with immunopositivity loss only after 8 days of ice storage [
35]. In this study, anti-calpastatin immunolabeling was maintained, while anti-dystrophin immunoreactivity disappeared [
35]. As in this European seabass study, dietary composition had a marginal impact on myofibrillar or endoprotease antibodies during the storage period. The different reactions of m-calpain and µ-calpain are related to their complementary functions and varying calcium requirements, which differ between fish species [
37,
114].
Under stress, such as starvation, cells can induce various strategies, including mitochondrial fusion or mitochondrial autophagy (mitophagy) [
116,
117]. This is compatible with the observations in the 65AS group under feeding restriction, where the ultrastructure study revealed elongated mitochondria involved in plastic processes. Mitochondrial fusion may increase under medium stress levels, providing a protective effect on autophagic turnover under nutrient deprivation conditions, inhibiting the onset of apoptosis [
118], and determining cell differentiation [
119]. Mitophagy combats stress by directly eliminating mitochondria, allowing the replacement of vital macromolecular precursors, including amino acids, sugars, and fatty acids [
116]. This result, first time described in fish, agrees with several studies in other species as mouse fibers, where mitochondria are significantly elongated and fused shortly after food restriction or starvation [
117,
118,
120].
Ice storage influences both the mitochondrial and fibrillar integrity. Mitochondrial dense granules, amplified at 4 dph, are intracellular accumulations of calcium deposits, generally in the form of calcium phosphate, within the mitochondrial matrix [
121]. The process of mitochondrial granule formation and mitochondrial membrane rupture may be interconnected phenomena. Under conditions of cellular stress, mitochondria accumulate excess calcium by forming dense granules; however, a calcium overload can induce mitochondrial permeability, resulting in the loss of mitochondrial membrane potential and the release of mitochondrial contents into the cytosol [
122]. This process activates proteases, such as calpains, which degrade structural proteins and contribute to the softening of fish flesh during the postmortem period [
114,
123]. These results are consistent with previous studies in GSB and European seabass, where dense granules and a decrease in mitochondrial membrane density were observed after approximately 4 days of ice storage [
33,
124,
125] increased after 6–7 days of ice storage.
Concerning the structural degradation of the fibers, at 1 dph, there was a slight detachment of the connective tissue from the fibers and an increase in interfibrillar separation. Rigor mortis is known to end after 12–36 h post-mortem due to the proteolytic activity of both calpains and lysosomal cathepsins, involved in the early degradation of muscle fibers observed in GSB [
33]. Previous findings on GSB and seabass reported that while no changes were observed in the myofibrils, part of the sarcolemma began to detach, and the interfibrillar spaces increased after 1 day of ice storage [
33,
126]. By 4 dph, these changes were found more frequently along the tissue. The alteration of the sarcomeres evidenced a loss of the integrity of the Z discs along with an increase in the amplitude of the I band. Some of these alterations could already be observed from day 3 in ice in the GSB muscle [
124], with significant structural damage identified in the fibers and myofilaments on days 5–6 of ice storage. These changes included fiber and endomysium detachment, as well as a decrease in the density of the Z discs and a barely identifiable I band [
33,
69,
124]. In European seabass muscle [
126], similar outcomes were described, whereas in rainbow trout, alterations were recorded at the fibrillar level after 7 days of ice storage, concomitant with the loss of fillet texture [
127]. Although proteolytic events appear to be common across different fish species, there are intra- and interspecific variations in the level of muscle degradation [
33,
126]. In addition, differences due to both intrinsic and extrinsic factors affecting enzymes and proteins [
38] are present.