The feeding trial and all the protocols conducted at the Curtin Aquatic Research Laboratory (CARL) were performed in strict compliance with the guidelines and regulations of Australia and the acts were reviewed and approved by the Curtin University Animal Ethics Committee (ARE2018-37).

The details of the diet formulations and fish rearing are reported in our earlier study (43). Briefly, four isonitrogenous and isolipidic were formulated: an FM-based diet control (0PBM-0HI) diet and three diets with the complete replacement of FM protein with 85% PBM and 15% HI larvae meal (85PBM-15HI), 80% PBM and 20% HI larvae (80PBM-20HI), and 75% PBM and 25% HI larvae (75PBM-25HI). The ingredients and chemical composition of the diets are presented in Table 1.

After acclimatization with the experimental condition, 300 barramundis with 7 g of mean initial weight were stocked into 12 tanks (73 × 84 cm and 300 L capacity) with 25 fish per tank. A biological filter, heater, and aerator were set up with each tank. Water quality parameters including dissolved oxygen, temperature, salinity, ammonia, nitrite and nitrate, and photoperiod were monitored daily and maintained within the recommended ranges as illustrated in our earlier study (22, 43). Each diet was assigned in triplicate and the fish were hand-fed twice a day (9.00 and 18.00) until satiety for 56 days. After 56 days of the feeding trial, the fish were not fed for 24 h and 8 fish/replicate were fileted immediately after stabbing the brain. Skinless filets from two fish were freeze-dried and stored at −80°C for further fatty acid analysis. For the remaining six fish, one side of the filets was transported immediately post-mortem on sterilized ice to a simulated display refrigerator for shelf-life study, while the other filets were subjected to individual quick freezing (IQF) with liquid nitrogen and stored at −80°C to preserve sensory quality until sensory evaluation.

The crude protein, crude fat, ash, and moisture were analyzed following the Association of Official Analytical Chemists (AOAC) (44) standard methods. For protein analysis, 1 g of the sample, a boiling chip, and a Kjeldhal catalyst tablet were placed in a digestion tube. Then, 8 ml of a sulfuric acid-phosphoric acid mixture along with 4 ml of 35% hydrogen peroxide were added to the tube, and the digestion was completed using a Kjeltec digester block (Foss Tecator 2020, Högänas, Sweden) at 420°C. After distilling the digested samples using the Kjeltec distilling unit (Foss Tecator 1002, Högänas, Sweden), the distillate was collected into a flask with 25 ml of boric acid indicator containing bromocresol green and methyl red which was subjected to titration with hydrochloric acid and conversion factor of 6.25 was used to calculate the crude protein content.

The crude fat content was determined using the petroleum ether extraction method where the fat from the samples was extracted using the Soxhlet unit (Extraction unit E-816, BÜCHI Labortechnik AG, Flawil, Switzerland). The extracted fat was then dried at 105°C till constant weight was obtained and the crude fat percentage was calculated by dividing the weight of the extracted fat over the sample weight.

For the ash content determination, the sample was weighed before and after heating overnight at 550°C using a muffle furnace (Thermolyne muffle furnace, model 48000, Thermo Fisher Scientific Inc, Iowa, USA). The ash content was calculated by dividing the post-heating sample weight over the initial sample weight.

For moisture estimation, the sample was weighed before and after oven drying at 105°C until a constant weight was obtained. The moisture content was then calculated by dividing the post-drying weight over the pre-drying weight.

The fatty acid of the experimental diets and whole freeze-dried filet (skinless) were analyzed following the protocol of O'Fallon et al. (45) as reported in our earlier study (28).

All the procedure-related sensory trial was carried out in strict compliance with the Australian Code for the Responsible Conduct of Research and National Statement on Ethical Conduct in Human Research, and reviewed and approved by the Curtin University Human Research Ethics Committee (Approval Number: HRE2020-0689).

The sensory quality of the barramundi filets at the end of the 56-day feeding trial was designated and evaluated following the method described by Gedarawatte et al. (46) and Lawless and Heymann (47). Eleven persons with no allergies, smoking habits, chronic health issues, visual impairment, respiratory issues, taste disorders, pregnant condition, not breast-feeding or on long-term medication, and who consume fish at least once every fortnight, were recruited, consented, and trained according to AS 2542.1.3:2014 (48) and CAC-GL 31-1999 (49). Before starting the sensory evaluation, screening on sensory sensitivity was conducted using multiple fish filets of various qualities. Three of the potential panelists were excluded due to their low precision in determining fish quality after repeat exposures. After that, 9 participants in the age of 18–50, consisting of 5 females and 4 males were screened and considered to be eligible and included in the study as semi-trained panelists. A labeled magnitude scale (50) was used to evaluate the visual appearance, odor, and overall acceptability of the raw barramundi samples in order to understand the quality of the barramundi in retail display.

Then, the same samples were sous-vided at 74°C for 6 min in vacuum-sealed boil-in pouches and served to the panelists within 15 min to evaluate the appearance, odor, texture, taste, and overall acceptability of the cooked samples similar to the raw samples.

The freshly fileted fish samples were immediately placed on ice inside an uncovered polystyrene box. The box was then placed in a 4°C refrigerator to simulate conditions in retail display. Melted ice was drained daily with new ice replaced. Six filets from different fishes were analyzed per treatment during each day on Days 1, 4, and 8 post-fileting.

The proximate composition of barramundi filet fed PBM-HI was similar to the 0PBM-0HI fed barramundi filet, except for crude lipid which increased in the filet of the barramundi fed with PBM-HI based diets (Table 2). The total saturated fatty acid (SFA) increased significantly due to a gradual increase in C12:0 (lauric acid) and C14:0 (myristic acid) in the filet of the barramundi fed with PBM-HI-based diets. A gradual increase of HI larvae meal in the diets also elevated the level C12:0 and C14:0 in the filets from the respective diets. An elevation of the C16:1n7 and C18:1cis + trans concentration resulted in an increase in the total concentration of monounsaturated fatty acid (MUFA) in the filet of the barramundi fed with PBM-HI-based diets. The lower concentration of total n-3 polyunsaturated fatty acid (PUFA) and n-6 PUFA, particularly, C20:5n3, C22:5n3, C22:6n3, C20:4n6, and C22:4n6, decreased the total PUFA in PBM-HI fed barramundi filets. A similar tendency was observed in the respective diets (Table 1).

The sensory evaluation of the raw and cooked flesh of the barramundi fed with fishmeal-free diets containing the different proportions of PBM and HI larvae meal for 56 days is presented in Table 3. PBM-HI-based diets improved the raw filet of barramundi, manifested by the higher scores given by the panelists for visual appearance, odor, and overall quality. Compared with 0PBM-0HI, the cooked odor was better in the filet of the barramundi fed with PBM-HI-based diets. The cooked texture and taste were unchanged by the test diets, whereas the cooked overall quality was better in the filet of the barramundi fed with PBM-HI-based diets when compared with the filet of the 0PBM-0HI fed barramundi.

Diet had no influence on drip loss while storage time had a significant effect on drip loss (Figure 1). The percent of drip loss increased significantly over the storage time irrespective of diet. There was no interaction between diet and storage time.

None of the texture profiles were influenced by the diet whereas storage time significantly influenced the cohesiveness, gumminess, and chewiness, as demonstrated by two-way ANOVA analysis (Table 4). Cohesiveness, gumminess, and chewiness decreased in all test diets over the storage time. There was no interaction found between diet and storage time in the texture profiles. Skin brightness and redness were significantly affected by diet. The brightness in the skin of barramundi fed with 85PBM-15HI and 80PBM-20HI decreased significantly on days 1 and 8. On Day 1, skin redness changed significantly in barramundi when fed with PBM-HI-based diets. Regardless of diet, skin redness over the 8 days storage time decreased. A similar change in the yellowness and chroma of the skin was found over the storage time, though diet did not change the skin yellowness and chroma. Filet brightness in response to test diets and storage time increased while flesh redness was not affected by diet and storage time. Flesh yellowness was increased by diet and storage time. Skin chroma was unaffected by diet and storage time. The microstructure of the muscle tissues in response to diets and over the storage time is presented in Figure 2. On Days 1 (Figures 2A–D) and 4 (Figures 2E–H), the muscle micro-structure was unchanged, supported by the tight attachment and regular shape of myofibrils together with a uniform distribution and distinct connective tissues. However, on day 8, the muscle tissue of the barramundi fed with 0PBM-0HI (Figure 2I) showed muscle degeneration and atrophy while the barramundi fed with PBM-HI-based diets (Figures 2J–L) demonstrated mild structural changes in muscles with 75PBM-25HI showing the least changes. It is also worth noting that an increase in intermyofibrillar spaces, reduced fiber-fiber adhesion, ruptured fiber, and loss of cell borders were observed on day 8.

Diet did not influence any of the attributes including skin brightness, appearance transparency, flesh texture, flesh blood, flesh odor, and flesh gaping used for QI schemes while storage time affected these attributes, as demonstrated by two-way ANOVA (Table 5). The score of all attributes increased over the storage time regardless of diet. No significant interaction was found between diet and storage time. There were a strong linear relationship between storage time and QI score in 0PBM-0HI (R² = 0.934) (Figure 3A), 85PBM-15HI (R² = 0.91) (Figure 3B), 80PBM-20HI (R² = 0.91) (Figure 3C), and 75PBM-15HI (R² = 0.91) (Figure 3D).

The results of the two-way ANOVA of the pH and TBARs activity of barramundi filet in response to diets and storage time are presented in Figure 4. The pH of the filet was influenced by diet and showed an interaction between diet and storage time (Figure 4A). The pH of the barramundi filet upsurged at Day 1 and 4 when fed with the mixture of PBM and HI larvae meal. However, storage time had no significant effect on pH. The malondialdehyde (MDA) content, measured by TBARS, was influenced by both factors with a significant interaction (Figure 4B). On day 8, PBM-HI-based diets retarded the production of MDA in the flesh of barramundi but there was a gradual increase in lipid oxidation production over the storage time irrespective of diets.

The crude protein levels were unaffected by the mixture of PBM and full-fat HI larvae meal, suggesting that barramundi can assimilate protein from PBM-HI as effectively as FM protein. Likewise, barramundi fed with poultry protein concentrate alone (6.7–20%) or supplemented with phosphorous (15), hybrid grouper fed blends of PBM, shrimp meal, and spray-dried blood meal (20–80%) (54) and juvenile black sea bass fed with PBM (40–100) (55) produced a similar whole-body protein content to fish fed FM protein. However, a higher lipid concentration reported here indicated the flesh quality improvement of barramundi flesh since lipid concentration has been reported to be associated with the taste and appearance of cooked flesh (56, 57). This result was mainly due to the inclusion of FHI larvae which contain a higher concentration of lipid.

Regarding the muscle FA profile, an increase in total SFA in the muscle of barramundi fed with 80PBM-20HI and 75PBM-25HI was due to the substantial increase in lauric acid (C12:0) and myristic acid (C14:0). The HI larvae meal is a rich source of C12:0 and C14:0 which have been reported to enhance the total SFA concentration in barramundi (23, 24) and other fish species such as Siberian sturgeon, Acipenser baerii (58), rainbow trout (59), and Eurasian perch, Perca fluviatilis (60). Oleic acid plus elaidic acid (C18: 1cis + trans), causing a higher concentration of MUFA in PBM-HI-based diets clearly improved the MUFA concentration of barramundi in the present study. This is similar to what has previously been reported in the MUFA concentration of barramundi (22–24, 28), juvenile coho salmon, Oncorhynchus kisutch (61), Atlantic Salmon, S. salar (62), and juvenile gilthead seabream, Sparus aurata (57) fed with PBM. PBM is a rich source of oleic acid, resulting in higher MUFA concentration in the aforementioned studies, however, it contains a lower concentration of n-3 LC-PUFA, EPA (20:5n-3), and DHA (22:6n-3). Similarly, a lower concentration of n-3 LC-PUFA, EPA, and DHA decreased the total PUFA in PBM-HI-based diets, perhaps underpinning the lower concentration PUFA in barramundi filet fed with PBM-HI based diets. Similarly, the high inclusion of PBM resulted in a lower concentration of PUFA in barramundi, totoaba juveniles, Totoaba macdonaldi, and juvenile coho salmon, Oncorhynchus kisutch.

Sensory quality is important to investigate the potentiality of new alternative protein ingredients and is associated with consumer acceptance. PBM-HI-based diets improved the raw and cooked filet quality of barramundi, perhaps suggesting that the aligned increase in filet lipid might be aligned with the improved sensory quality of the barramundi filet. Besides lipid content, chitin present in HI larvae might have played a role in improving sensory filet quality as chitosan coating has been reported to improve the sensory scores in seafood (63–66), however, the mode of actions of chitin in HI larvae meal after feeding in improving the filet quality is not well-understood. The present study provides the first insight on the improvement of barramundi filet quality when fed with FM–free diets containing a mixture of PBM and HI larvae meal. However, the total substitution of FM with PBM deteriorated the sensory quality of female tenches, Tinca MacDonald (26), and meat meal did not affect the sensory quality of the barramundi filet (25). Hence, further study is needed to identify the components in HI larvae meals that may improve the filet quality of fish.

Evaluation of drip loss is of importance in determining the post-harvest fish quality due to the fact that it is directly related to lower water holding capacity caused by fiber shrinkage, cell damage, lower protein solubility, and protein denaturation (67), leading to sensory loss including texture and flavor (68), loss of water-soluble nutrients such as protein and amino acids (69), and financial implications (68). The unchanged drip loss in the filet of barramundi fed PBM-HI based diets similar to the unchanged filet water holding capacity of rainbow trout fed HI larvae prepupae meal (70) and blackspot sea bream, Pagellus bogaraveo fed mealworm larvae meal (71). This finding is consistent with the results of texture and lipid oxidation. Regardless of diet, gradual increase in drip loss over the 8 days storage is consistent with the findings of Siah and Ariff (72) who reported an increase in drip loss in barramundi flesh samples over time during storage at 2 ± 2°C but was not consistent with the findings of Wilkinson et al. (73) who reported no differences in drip loss in barramundi flesh samples over 4 days storage at 2–4°C. The result is within expectation as degradation in structure and gradual increase in lipid oxidation was found, indicating that the integrity of cell membranes reduces with storage time, enhancing passage of sarcoplasmic fluid which lead to an increase in drip loss (74, 75). However, the changes in drip loss from 0.16 to 2.86% cannot be considered as high and therefore not a major problem in chilled barramundi, as reported by earlier studies (67, 76, 77).

A good number of studies have been evaluated the efficacy of FM replacement with PBM on 33 different fish species, though none of them have been evaluated the flesh quality (20). However, our recent study found no changes in the final flesh quality of barramundi in terms of pH, texture, and color when fed PBM-based diets, concurrently supplemented with tuna hydrolysate and HI larvae meal (24). Similarly, in this study texture including springiness, cohesiveness, gumminess, chewiness, and hardness of barramundi filets were not affected by the mixture of PBM and full-fat HI larvae meal. Regardless of PBM inclusion, HI larvae meal (33–100%) had no influence on the texture (hardness, cohesiveness, resilience, and adhesiveness) of salmon (78). This is similar to what has been reported by Iaconisi et al. (71) who found no variation in filet hardness, cohesiveness, gumminess, and adhesiveness of sea bream, Pagellus bogaraveo fed mealworm, Tenebrio molitor. However, storage time had a significant effect on the cohesiveness, gumminess, chewiness, and hardness, which decreased significantly on day 8. A similar result was found in shelf-life studies conducted in sea bream, Sparus aurata, where identical parameters reduced over time (79, 80). This phenomenon has shown to be highly related to histological structure, especially the increase in intermyofibrillar spaces, reduced fiber-fiber adhesion, ruptured fiber, and loss of cell borders. Thus, it can be suggested that textural changes were a consequence of structural degradation over time. These results are further supported by elevated lipid oxidation over the storage time in the present study, suggesting that protein deterioration took place by the action of endogenous cathepsins and exogenous protease due to microorganisms (81, 82). Hence, whole microbiome profile evaluation using modern tool rather than plate counting in barramundi filet in response to PBM-HI based diet during shelf life study is needed to be thoroughly investigated.

The Color of fish is an important factor influencing consumer preference and it is largely governed by dietary and post-harvest factors in fish since fish do not have de novo power to synthesis color. There is a lack of studies concerning the effect of processed animal protein on the color of fish. However, the complete replacement of FM with PBM, concurrently supplemented with HI larvae meal and tuna hydrolysates did not change skin and filet color in barramundi (24). In this study, the skin brightness (L^*) was affected by diet and this variation in the present study could be due to light or lack of shading. Farmed fish are more exposed to light than wild fish, leading to darker skin chromophores. These suggest the farmed fish may be darker than wild fish (83). This was proven by Howieson et al. (16) who reported that barramundi reared in shaded tanks were less dark than the barramundi reared in unshaded tanks. PBM-HI-based diets improved the redness (a^*) in barramundi skin which could be due to the presence of β-carotene in insects (84). A similar effect was observed in blackspot seabream, Pagellus bogaraveo skin when fed mealworm, Tenebrio molitor (85).

Blue-gray discoloration due to melanosis is one of the common problems in farmed barramundi filet, reducing the appeal of the raw farmed barramundi filet on display to consumers and consequently impacting sales in the retail sectors (16). The same authors fed barramundi for 6 weeks by changing the levels of the substrate (the amino acid tyrosine) or substrate competitors (the amino acid tryptophan), or limiting enzyme (tyrosinase) co-factors like copper in the experimental diets to manipulate the melanin synthesis pathway, and however, none of these diets improved the graying issue. Irrespective of storage time, the improvement in barramundi filet brightness when fed with PBM-HI in the present study suggested that HI larvae meal might have properties to reduce melanosis. This is not similar to the report of Bruni et al. (70) and Moutinho et al. (86) who found no differences in the filet brightness of European seabass and rainbow trout fed with HI prepupae meal. However, brightness over the 8 days storage time increased irrespective of diet which is similar to the brightness of barramundi flesh over the 14 days of storage time (12). Also, the brightening of fish flesh over the prolonged storage time has been frequently reported by many studies (87–90). It could be due to the elevated lipid oxidation during the storage time which has been reported to affect the flesh color of fish. Flesh brightening may also be attributed to some other factors including alterations in light scattering properties in muscle during rigor (90, 91) and light-absorbing and reflecting properties due to decreases in the translucency of fish muscle (92, 93). Riboflavin (vitamin B2), a yellow-colored pigment variably found in most edible insects (71), may also influence the skin and flesh color of fish. The improvement in yellowness (b^*) in the flesh of barramundi fed with PBM-HI-based diets could be due to the presence of 2.2 mg/100 g riboflavin in the HI larvae meal (94). In contrast, a 6-week feeding trial on barramundi fed with PBM-based diets supplemented with 5 and/or 10% of HI larvae and tuna hydrolysate did not influence the skin and flesh yellowness (24). A similar influence to a previous study on the yellowness of European seabass, Dicentrarchus labrax juveniles filets was found when fed pre-pupae larvae (86). These discrepancies in color variation might also be attributed to different inclusion levels of HI larvae meal, trial duration, and different fish species.

Quality index (QI) is a scoring system to estimate the freshness and quality of fishery products based on some attributes including appearance, odor, and texture changing during the storage time. The score for all attributes increased gradually during the storage time but the score was below 3 (low freshness) (52) for all attributes, indicating that barramundi filets were acceptable until day 8. However, a significant linear relationship between QI score and storage time was expected since a similar trend has been reported in many fish species (95–100). After 8 days of storage, the highest demerit score for QI for all diets was similar to the results of textural attributes, muscle microstructure, and lipid oxidation in the present study.

The final product quality deterioration of fish including color degradation, muscle gaping, blood spotting, flesh texture alterations, and drip loss is associated with the reduction of pH (12, 73, 101). An elevated level of pH in the filet of barramundi fed with PBM-HI based diet in the present study is similar to what has been reported by Moutinho et al. (86) in which pH increased in the filet of European seabass fed pre-pupae meal after slaughtering and 3 days post-storage. However, pH over the storage time in the present study changed from 6.34 to 6.74 was within the normal pH range of barramundi muscle (73).

Although pH elevation may be attributed to the formation of nitrogenous compounds such as ammonia, dimethylamine, trimethylamine, histamine, etc., produced by endogenous enzymes and microbial enzymatic actions (102), the increase in pH at certain levels has been reported to prevent the production of volatile bases, leading bacteria to take energy from the oxidative product rather than glycogen and other normal substrates (12). Hence, an inverse association between pH and lipid has been reported in barramundi filet after 5 months of feeding of alpha-tocopherol acetate and following a 2-week post-storage time (12). A similar association was found between pH and TBARS analysis in the present study. TBARS is widely used to quantify the degree of second stage lipid peroxidation of fish that produce aldehydes and ketones from the degradation of polyunsaturated fatty acids, associated with unpleasant odor (103). PBM-HI-based diets inhibited the rancidity production (MDA) compared with 0PBM-HI in the current study, the reduction in MDA level could be related to elevated MCFAs level as it reduced the amount of double bond which is prone to be attacked by an oxidant (104). A similar trend was also supported by Table 2 fatty acid profile, where a decrease in TBARS was only found when the MCFAs were significantly higher. A previous study has suggested that HI larvae may possess antioxidant properties to halt lipid oxidation. For example, Moutinho et al. (86) reported the efficacy of HI larvae meal to reduce the production of TBARs in the filet of European seabass. The less production of TBARS in the present study might be attributed to the lower proportion of PUFA, in particular, EPA and DHA which are sensitive to peroxidation in contact with oxygen (105).

Or, the presence of chitin in HI larvae might have retarded oxidation since the deacetylated form of chitin is known as chitosan which has been reported to delay lipid oxidation by binding the free amino groups and hydroxyl radicals of the polymer with the metal ions (Fe²⁺) and free radicals on food, thus making stable macromolecular structures (66, 106, 107). Another potential reason behind these improvements in rancidity is associated with the interaction between positively charged (NH3+) amino groups of chitosan with the negative carboxyl groups (COO⁻) situated on the outer part of the membrane or the cell wall of bacteria and fungi (66), which change the permeability of the membrane of by blocking the passage of nutrients and oxygen important for cellular metabolism, leading to cell death (108, 109). Although chitinolytic enzymes, chitinoclastic bacteria, and the assimilation of chitosan have not been reported previously in barramundi, endogenous production of chitinase has been previously reported in marine carnivorous teleost fish (110, 111). Regardless of the diets, the flesh TBARS increase over the storage time in the current study is consistent with the reported TBARS production of barramundi flesh during 2-weeks chilled storage (12) and in filets of European seabass when fed with different levels of HI pre-pupae larvae meal (86). The range (0.70–6.69 mg MDA/kg) of TBARs production in the present study up to Day 8 was below the reported critical limits (7–8 mg MDA/kg) for fish (112, 113).