Protein Hydrolysates from Fisheries Processing Byproducts: Production, Characteristics, Food Applications and Challenges

1. Introduction

In response to the recognition of the limits of marine sources and the crisis of food security, production of aquatic food products is increasing and reached 179 million tonnes in 2018, of which 22 million tonnes (or 12%) were not used for human consumption [1]. Maximum use of by-products from fish processing industries (e.g., heads, frames, skin, trimmings, and viscera from fish, cephalothorax and shells from shrimp, heads and tentacles from squids, and shells and byssus threads from oysters and mussels), that account for 40-60% of the weight, are necessary to retain them in the food chain in line with a sustainable circular economy through production of high-value biomolecules [2,3,4,5]. Among the different fractions of by-products that are produced after filleting, canning, packaging, etc., heads, frames and trimmings constitute >75% of the by-products’ weight and contain significant amounts of muscle residue that can be used for direct human consumption or can be converted to functional food ingredients. Despite the necessity, using by-products within the seafood sector is happening relatively slowly because the seafood industry is more focused on its primary raw materials and products that require minimal processing [6].

One focus has been peptides from marine sources (fish, shellfish, and invertebrates) with antioxidant activity [7,8,9,10]. Antioxidant peptides have an important role in inhibiting oxidation and scavenging free radicals. In the body they may help fight aging and reduce food oxidation. These peptides have been prepared using enzymatic or autolytic hydrolysis, or microbial fermentation [2]. Most peptides studied have been 2-10 amino acids (AA), although some were up to 20 AA and had molecular weight (MW) of 0.2 to 2 kDa [11,12,13,14]. Several proteases with different specificity for peptide bond cleavage and different hydrolysis conditions (temperature, time, pH, enzyme to substrate ratio, water/by-products ratio, and stirring rate) have been used to produce these marine by-products peptides [15,16,17,18]. Changes of hydrolysis conditions did change their antioxidant activities [19,20,21]. Antioxidant activity of peptides in foods has been associated with scavenging of the free radicals formed during lipid peroxidation, and metal chelation [22], which are dependent on the AA composition and sequence, size, hydrophobicity, and N- or C-terminal residues [23,24,25].

Marine protein hydrolysates derived using enzymatic hydrolysis showed antioxidant activity against free radicals and pro-oxidative metal ions. Thus, they may potentially be used as alternative antioxidants in foods and the human body to fight against free radicals-mediated aging [26,27]. The structure of antioxidant peptides and protein hydrolysates from marine sources are highly variable depending on the types of by-products used as the initial protein source as well as the various operating parameters that affect the hydrolysis and the functional outcome. These by-products might be stabilized to preserve their freshness with no oxidized products to ensure lower deteriorative reactions during enzymatic hydrolysis which is needed to obtain protein hydrolysates with acceptable organoleptic properties and storage stability. Marine protein hydrolysates and peptides have been shown to reduce oxidation of both lipids and proteins of seafood during storage, thus indicating antioxidant and anti-freezing effects [21,28,29,30].

Despite the suggested applications for protein hydrolysates from by-products in the food industry, there are several challenges that have caused such products to be unable to be manufactured on a commercial scale. This review will try to discuss the production and characteristics of protein hydrolysates from fisheries by-products derived using hydrolysis, factors involved, and their potential applications to control/reduce oxidative deteriorations of seafood during storage. It also reviews the structure-activity relationships of the antioxidant activity of peptides as well as challenges of industrial processing of by-products and commercialization of protein hydrolysates production.

2. Fish Protein Hydrolysates: Production and Processing Factors

The fish protein hydrolysates markets are anticipated to reach USD 558 million by 2025 with compound annual growth rate (CAGR) of over 5% due to increasing demand for protein-based supplements, food formulations, infant nutrition, fertilizers, and aquafeeds due to their higher absorption or digestion [2]. These hydrolysates contain peptides with 2-10 amino acids or up to 20 amino acids and molecular weight of <3 kDa, especially between 0.2 to 2 kDa [2,14]. By-products of several commercial species such as tuna [17,20,31], tilapia [12,32,33,34,35], marine bony species [19,36,37], small pelagics [38,39,40], salmonids [15,16,41,42], shrimp [18,43,44], marine invertebrates such as mollusks [45,46,47], squid and cuttlefish [48,49,50,51,52,53], and underutilized fish [54,55] have been used as the source for producing protein hydrolysates. By-products from those species showed differences in their composition (i.e., different amounts of lipids, blood, proteins, undigested feed in their stomach or intestines, etc.), emphasizing the need to adjust the hydrolysis conditions for each source according to the inherent characteristics of those by-products (such as sorting by-products to obtain different fractions). Depending on the farm conditions or ocean water quality, there are some safety issues with by-products from different category of aquatic species that should be considered as undesirable metabolites and contaminants that may enter the liquid phase in which peptides are formed, affecting the safety of the protein hydrolysates. Table 1 shows the yield, composition, safety concerns, and preventive measures for by-products from the main groups of aquatic species (fish, crustaceans, and invertebrates).

Enzymatic and autolytic hydrolysis and microbial fermentation have been used for converting the proteins of by-products into peptides with varying sizes and bioactivities [2,7,8,78]. The process of hydrolysis with commercial proteases can be controlled by selecting appropriate enzymes and adjusting the hydrolysis conditions (time, temperature, enzyme concentrations, water ratio, etc.), and the final hydrolysates should be stable in terms of structure (such as peptide profile) and functions. Autolytic hydrolysis mediated by endogenous enzymes such as acid/aspartyl proteases (such as pepsin), serine proteases (trypsin and chymotrypsin), thiol/cysteine protease (cathepsin B, L, and S), and muscle proteases (lysosomal cathepsin, alkaline and neutral protease) have been used to hydrolyze marine proteins [79,80]. However, acidic proteases have a lesser role in autolysis because most studies have shown that hydrolysis by endogenous proteases usually occur at around neutral pH [2] unless this group of proteases is first isolated and then their effects are investigated in acidic pH [81,82,83], although, the cost associated with purification of viscera proteases, their stability, and activity may be a challenge compared to highly stable commercial enzymes. With autolysis, several variables such as the freshness of by-products may affect optimal proteolytic activity and the possible denaturation of endogenous enzymes due to improper handling and storage of by-products will affect the autolysis efficiency and thus, the end products’ properties. But because autolysis is not part of the goals of this review, more detailed information on autolytic hydrolysis technology can be found in previous studies [2]. In general, to obtain protein hydrolysates with stable peptide profiles, smell, nutritional value, functional properties, biological activities, and bioavailability, the process of enzymatic hydrolysis must be controlled at each step of the production process. Figure 1 shows the factors involved and main issues to be considered during enzymatic hydrolysis of fisheries by-products. Properties and structure of the end products are governed by three main factors (the by-products, enzymes, and operating parameters) (Subsections 3.1-3.3).

2.1. By-products Composition, Quality, Storage and Handling

The type of by-products (viscera, heads, frames, trimmings, or their mixtures) and their composition (different amounts of lipids, blood and hemoglobin and metal ions as pro-oxidants) influence the composition, nutritional, and antioxidant properties of the protein hydrolysates [15,84]. Frames, trimmings and heads are clean by-products with potential uses for direct human consumption or as functional hydrolysates. A recent study has shown that 37.2, 56.7, and 81.0% of the weight of heads, frames, and trimmings of Atlantic salmon, respectively, are edible as direct human food while for the viscera, the edible yield is normally thought of as 0% [10], due to the presence of lipids, bile acids, blood, non-digested feed in the intestine, etc., associated with a high rate of oxidation and formation of undesirable metabolites during hydrolysis [3]. According to a survey of methods to utilized Scottish salmon by-products, 15% of the final utilization (frames, trimmings, heads) is for food, 75% (frames, heads, viscera, mixed by-products, skins) is for feed purposes, and 10% (blood) is for fuel and fertilizer [6]. Therefore, cleaner by-products (frames, heads, trimmings) with acceptable freshness are ideal for food pouposes, including production of protein hydrolysates while viscera (alone) is used for producing hydrolyzed protein concentrate and oil, or fish meal and oil (rendering) in the form of mixed by-products. Some of the oxidation products can add carbonyl derivative to the peptides, decreasing their antioxidant activity and possibly their safety. In this regard, Aspevic et al. [85] reported the differences in essential AA, biogenic amines, and sensory properties of protein hydrolysates from backbones, heads, and viscera of salmon and mackerel. Viscera hydrolysates from salmon (SV) and mackerel (MV) had more intense taste and bitterness compared to hydrolysates from heads and backbones. They also showed more small peptides with MW <200 Da (64.7 and 74.5% for SV and MV, respectively) due to the activity of endogamous enzymes during enzymatic hydrolysis with FoodPro PNL (10 U/g protein, 50 min at 55 °C). The presence of bile may add to the bitter taste sensation. SV and MV had higher TMA, cadaverine and putrescine compared to hydrolysates from heads and backbones. MV showed more taste intense and had higher scores for umami, salty, and fish taste compared to salmon hydrolysates, partly due to a higher content of ash (38.3%) in mackerel heads hydrolysates, indicating the need for salt removal before human use. Heads and backbones hydrolysates showed much lower free AA than viscera hydrolysates in both species (~40 g/100 g visceral protein), indicating autolytic hydrolysis that causes higher free AA when using viscera as the raw material for protein hydrolysates. Despite these results, viscera hydrolysates are unlikely to be suitable for use in foods due to safety issues. On the other hand, its potential as functional feed ingredients in the form of acid silage or liquid hydrolysates can be used to enhance the growth, nutrition and health at the larval stage of aquatic species [86]. However, in most studies, the purpose of hydrolyzing viscera and identifying its antioxidant peptides is not clearly stated, although it is generally suggesting the direct use of viscera hydrolysates (which raises safety issues) in foods, or the use of synthetized peptides, which need to consider the economic viability.

The freshness of by-products is one factor affecting the formation of oxidative compounds during hydrolysis, and eventually the structures, functions and shelf life of protein hydrolysates [2,3]. Endogenous muscle proteases including matrix metalloprotease (MMP), and serine and cysteine proteases degrade myofibrillar proteins and microfibrillar networks, resultuing in a significant decrease of quality [87–89]. These proteases are active in trimmings, frames, and heads, which with inappropriate storage temperatures, degrade myofibrillar proteins, lowering the initial quality of proteins for further hydrolysis. A recent survey of fisheries by-products handling practices in Europe indicated that the sorting of by-product fractions was done in >60% of the seafood processing plants, while the remainder did not handle their by-products properly and only 25% of the surveyed plants managed by-products as food grade. According to that survey, the majority of the processing companies used their by-products for non-food purposes, mainly as feed [58]. A study on salmon by-products (frames, heads, and viscera) showed that storage at 4 or 10 ºC greatly influenced the quality. Several metabolites including tyramine, histamine, trimethylamine (TMA) have been formed, and the decline in the quality of viscera increased histamine level that exceeded the limits after 2 days of storage at 10 ºC while, those by-products could be stored for up to 7 days at 4 ºC [57]. Therefore, by-products processing should be done near the fish processing centers so the hydrolysis of by-products can be done in the shortest possible time. When the quantity of by-products exceeds the capacity of the plant, they should be stored frozen. Therefore, the impact of frozen storage for varying times needs further study on the structure of peptides (AA composition, sequence, and size) and their biological functions.

Pretreatments of by-products may lower the concentration of pro-oxidants while resulting in purer protein substrates for enzymatic hydrolysis. Rinsing or incubation of herring by-products (heads, backbone with caudal fin, skin, intestines, and eggs) with antioxidant solutions including Duralox-MANC, isoascorbic acid, isoascorbic acid + ethylenediaminetetraacetic acid (EDTA)) decreased the rate of lipid oxidation and hemoglobin levels during storage at 4ºC for up to 12 days while, extending the shelf-life of by-products from <1 to >12 days with a rinsing strategy or to >7 days with direct addtion into the by-products after mincing [62]. However, due to the high rate of lipid oxidation in such a highly sensitive system, upgrading to food grade with proper safety is a big challenge. Washing or defatting of underutilized Sind sardine muscle mince significantly decreased total pigments (600 and 140 μg/g dry sample) and heme iron (5.3 and 1.2 mg/100 g dry sample) content in washed and defatted mince substrates, respectively, compared with non-pretreated mince (2570 μg/g dry sample and 23 mg/100 g dry sample for total pigments and heme iron, respectively), resulting in lower formation of TBARS but increased DPPH radical scavenging and ferrous chelating activities of protein hydrolysates, espcially with defatted mince [54]. Pretreatment of cape hake by-products with 8 mM CaCl₂ + 5 mM citric acid followed by alkaline solubilization of proteins (pH 11) resulted in significantly lower phospholipids (1 AU/g) and lipids (0.39%) but higher solublity in hydrolysates compared to samples directly produced from by-products (4.2 AU/g and 0.87%, respectively). However, these hydrolysates were characterized by higher yellowness and redness that was attributed to the alkaline solubilization of heme proteins during protein isolation [90]. Despite the relative improvement in by-products quality, the economic issue of carrying out such pretreatments at the tonnage scale (i.e., industrially) is a challenge that has not been studied. To tackle technological problems associated with heme proteins and to lower oxidation, the use of antioxidative extracts from agricultural wastes including lingonberry press-cake, apple-, oat-, barley- and shrimp by-products and seaweed (ulva) extracts as helpers all at 30% of the dry weight of the by-product, decreased formation of MDA and the oxidation product 4-hydroxy-(E)-2-hexenal (HHE) in herring and salmon heads and backbone protein isolates, resulting in more stable substrates. Among all helpers, lingonberry press-cake followed by apple peel and ulva were the most effective in reduction of lipid oxidation during alkaline solubilization/acid precipitation and 9 days of ice storage. The new color (dark purple) in the resulting protein isolates with lingonberry press-cake might be advantageous for increasing acceptance of the color by consumers [91]. Further work is needed on the interactions between the phenolic compounds of plant extracts as antioxidants and the by-products proteins, and how it will affect the structure and biological activity of the protein hydrolysates.

2.2. Proteolytic Enzymes

Composition and sequence of peptides in whole hydrolysates from the same source of protein may differ depending on the type of enzyme used. Endopeptidases (e.g., trypsin, chymotrypsin, pepsin, pancreatin, papain, and Alcalase) act away from the N- or C-terminus, while exopeptidases (e.g., carboxypeptidase Y, aminopeptidase M, and Flavourzyme) break peptide bonds at the terminus of polypeptide chains [92]. Since different enzymes have specific cleavage sites (papain: Arg-, Lys- and Phe-; Alcalase: Ala-, Leu-, Val-, Tyr-, Phe- and Try-; trypsin: Arg- and Lys-; pepsin: Phe- and Leu-), different cleavage sites will affect the AA composition, and sequence of peptides [93]. Salmon skin gelatin hydrolyzed with Alcalase showed a higher content of hydrophobic AA, degree of hydrolysis (DH), surface hydrophobicity and peptides with MW <1 kDa than hydrolysates produced using Neutrase, Protamex, and Flavourzyme. This was associated with significantly higher OH^• and O₂^•- scavenging and Fe²⁺ chelating activity [42]. From tilapia skin gelatin, different peptides, including Gly-Pro-Ala [12], Glu-Gly-Leu (317 Da) and Tyr-Gly-Asp-Glu-Tyr (645 Da) [32], Asp-Pro-Ala-Leu-Ala-Thr-Glu-Pro-Asp-Pro-Met-Pro-Phe (1383 Da) [33], Leu-Ser-Gly-Tyr-Gly-Pro (592 Da) [34] and Tyr-Gly-Thr-Gly-Leu (509 Da) and Leu-Val-Phe-Leu (490 Da) [35] were obtained depending on the enzyme used despite starting with the same protein (tilapia skin gelatin (although the method of production of the gelatin may have differed)). In abalone viscera, different enzymes resulted in different peptide sequences; Alcalase: Gln-Ser-Cys-Ala-Arg-Phe (711 Da), Ala-Ala-Pro-Ala-Val-Ser-Gly-Arg (728 Da), Asn-Arg-Phe- Gly-Val-Ser-Arg (834 Da), and Pro-Val-Pro-Pro-Tyr-Lys-Ala (770 Da), Neutrase: Ala-Ala-Gln-Tyr-Ser-Arg-Asn (808 Da), Val-His-Ala-Glu-Pro-Thr-Lys (780 Da), Gly-Cys-Tyr-Val-Pro-Lys-Cys (769 Da), and Asn-Ser-His-Val-Val-Arg (711 Da); papain: Ala-Ala-Asn-Asn-Ser-Thr-Arg (732 Da), Thr-Ile-Asp-Cys-Asp-Arg (722 Da), Cys-Ile-Gly-Tyr-Asp-Arg (725 Da), Asp-Asp-Ile-Thr-Arg-Asp (734 Da), and Asp-Val-Ala-Phe-Met-Arg (738.3 Da); and trypsin: Met-Glu-Thr-Tyr (543.3 Da), Tyr-His-Gly-Phe (523 Da), Gln-Cys-Val-Arg (505 Da) [45]. Tyr-Pro-Pro-Ala-Lys (574 Da) [46] and Pro-Ile-Ile-Ser-Val-Tyr-Trp-Lys (1005 Da) [47] have been purified from blue mussels using Neutrase and pepsin, respectively. Despite the influence of the structure of peptides on the selectivity of the enzymes used, the high cost of commercial proteases suggests the minimal use of enzymes for hydrolysis. As with Atlantic cod, hydrolysis of heads with a combination of papain and bromelain at minimal concentration of 0.1% by weight of the minced heads for 1 h resulted in reasonable peptides profile in which 56% of peptides were <2 kDa, while most peptides (~33%) had MW between 1 and 2 kDa [94].

2.3. Operating Parameters

The temperature and pH are generally adjusted according to the selected proteases to ensure high hydrolytic activity. Thus, the other factors such as the water to by-products ratio, type of propeller, stirring rate, enzyme deactivation step, use of nitrogen gas, antioxidant addition, time, etc. should be optimized to ensure a stable end-product.

Abalone food muscle hydrolyzed with papain (HPP), Protamex^® (HP) or an animal protease (HA) for 0.5 or 4 h showed differences in physicochemical and structural properties governed by the time of hydrolysis and enzyme type. The fluorescence emission spectra of all hydrolysates showed a red shift of 10-12 nm compared with that of control while, fluorescence intensity was higher in hydrolysates than non-hydrolyzed proteins (AFP). Hydrolysis for 4 h resulted in higher intensities in HPP, HP, and HA compared to lower hydrolysis time. Among all samples, HA-4, HPP-0.5, and HPP-4 had higher absolute ζ-potential values than AFP, indicating a higher number of ionizable groups on the protein surface that inhibited the formation of protein aggregates, consistant with solubility, S-S bonds and free –SH groups [95].

de la Fuente et al. [96] reported the differences in peptides identified from salmon viscera obtained using two different stirring methods including conventional stirring (30 min in distilled water at room temperature) and pressurized liquid extraction (PLE; 1500 psi, distilled water as solvent, pH 7.0, 50 ºC, 15 min) despite using the same protein source. For viscera subjected to PLE, 137 peptides were identified and contained several small antioxidant peptides with sequences of Gly-Pro-Pro and Gly-Ala-Ala. On the other hand, only 67 peptides were identified in the control extract. In both extracts, the MW was in the range of 0.6 to 2.6 kDa. However, the PLE extracts contained greater amounts of small peptides. Total antioxidant capacity, measured using ORAC and TEAC showed that PLE viscera extract had higher values (3790 and 7770 μM Trolox Eq for TEAC and ORAC, respectively) compared to the control extracts (788 and 2450 μM Trolox Eq for TEAC and ORAC, respectively).

During hydrolysis, especially at the industrial scale, the high water addition increases the production costs due to the heating and drying needed to obtain a powder or a concentrated liquid [3]. Using less water during hydrolysis can be beneficial as long as it does not affect other processes and efficiency. For cod head hydrolysates with different water to heads ratios (1:1, 1:0.75, and 1:0.5 kg/kg), the ratio was found to have little effect on hydrolysis yield, protein content, and MW distribution of peptides and thus, high water addition might possibly be unnecessary. In addition, the hydrolysis reaction time of 1 h was suitable to obtain hydrolysates with desirable properties [94].

Lipid oxidation is one of the main challenges during enzymatic/autolytic hydrolysis of by-products, resulting in unpleasant odors and flavors, dark colorations and formation of oxidative products in the FPH [2,3]. Due to the low lipid content of heads (1-4%), a separated oil faction was not formed in the cod head hydrolysates (0.65% lipids in the FPH) [94]. When working with a mixture of cod viscera and trimmings, a separate oil faction was formed after enzymatic hydrolysis and a minimum of 6 g of lipids/100 g wet weight by-products was required to form an oil fraction [97]. The intensity of lipid oxidation during hydrolysis will be different among by-products with different compositions, which may affect the structure and safety of the resulting peptides and their antioxidant activity. Reduction of oxygen and replacing it with an inert gas such as nitrogen were shown to decrease lipid oxidation in Sind sardine [54] and tuna [98] protein hydrolysates.

Enzyme deactivation is the last step of the hydrolysis process. The high temperature (80-100 ºC, 10-15 min) is often used to deactivate enzymes [11,12,13,14,15,16]. This temperature may lead to structural changes of the protein hydrolysates and peptides, decreasing antioxidant or other bioactivities. Xie et al. [99] used slow (+4 ºC) and rapid (-18 ºC) cold deactivation temperatures, a 100 ºC water bath, and no deactivation and found that the deactivation method significantly affected the DH, surface hydrophobicity, average particle size, intrinsic fluorescence, secondary structure content (α-helix, β-sheet, β-turns, and random coils) and antioxidant activity.

Storage of protein hydrolysates as a powder following spray-drying may affect their function and properties during storage due to hygroscopicity. If stored at temperatures above the glass transition temperature (T_g), protein hydrolysates will be sticky (10 – 20 °C above Tg), cake (20 – 30 °C above T_g) and collapse (40 – 50 °C above T_g). Thus, to avoid loss of quality and function, powders must be stored at temperatures below their T_g [40].

2.4. Process Scale-Up

One roadblock towards the industrial production of protein hydrolysates from by-products is that most studies with by-products were done at the laboratory scale, which limits their industrial adaptation [114]. Some studies attempted pilot trials to confirm the technical feasibility of the laboratory scale at an industrial scale. In this sense, the process of producing peptides from hake by-catches was scaled up from a 0.5 to a 150 L reactor using the optimized hydrolysis conditions (2% enzyme, 2 h, 50% solids, pH 9, 70 ºC) that were obtained in the laboratory [101]. The authors found similar results at the pilot plant scale in term of protein extraction yield (60.0% pilot and 61.4% lab), antioxidant capacity (172 mg TEAC/g protein in pilot and 224 mg TEAC/g protein in lab), and antioxidant capacity yield (103 mg TEAC/g protein in pilot and 132 mg TEAC/g protein in lab). Furthermore, liquid, solid, and bone yield did not show any significant differences from the results of the laboratory trials. Monkfish by-products (heads and viscera) hydrolysis was scaled up from 100 mL to a 5 L glass reactor at optimized laboratory conditions: 57 ºC, pH 8.3, solid to liquid (S/L) ratio of 1:1 (w/w), 0.05% Alcalase and 200 rpm stirring rate for 3 h. Following hydrolysis, the hydrolysates were filtered (100 μm) to remove non-hydrolyzed materials. Results validated the properties of the FPH obtained by the optimization trial at the laboratory scale. However, the 5 L reactor may still be considered a bench-scale trial and it needs work with larger reactors [110]. An industrial-scale production of hydrolysates from Atlantic salmon by-products (heads, frames, and viscera) using the laboratory (4 L glass vessel) hydrolysis parameters such as enzyme type and concentrations, time and temperature, except for a shorter time from slaughter to hydrolysis (2 h), avoidance of preheating of water inside the reactor, separation using a decanter, and no drying of the hydrolysis solution (acidified with formic acid to reach pH <4) was tested in an industrial plant. The authors attributed the main differences to lower hydrolysis efficiency, separation, and storage conditions [15]. Overall, the cost of enzymes, high water usage, drying or condensation of the hydrolysis solution, as well as the use of separators (one or more steps) to remove oil and other undigested materials are the main cause of the greater process costs at an industrial scale, although the complexity of the hydrolysis line depends on the type and composition of the by-products.

3. Antioxidant Activity of Fisheries By-Products Protein Hydrolysates and Peptides

The antioxidant peptides had 2-10 amino acids, although some were up to 20 amino acids and had a MW of 0.2 to 2 kDa. Antioxidant activity of peptides is mostly related to the presence and position of specific amino acid residues in the peptide chain. Primary structure, amino acid composition, hydrophobicity, spatial conformation, etc. are characteristics that are affected by enzymatic hydrolysis and determine its antioxidant activity [115]. Peptides show antioxidant activity due to the presence of one or more hydrophobic (Pro, Ala, Gly, Leu, Ile, Met, Trp, Phe, Val) and aromatic (Tyr, Trp, Phe) amino acids that can quench free radicals by various mechanisms including hydrogen atom transfer (HAT), single electron transfer followed by proton transfer (SET-PT), and sequential proton loss by electron transfer (SPLET) mechanisms [116]. Scavenging free radicals and oxidants using HAT, SET-PT, and SPLET generally leads to the same end results, although the kinetics and potential for side reactions vary. SPLET, SET-PT, and HAT may occur in parallel; however, the dominant mechanism depends on the antioxidant’s conformational and geometrical features, solubility, partition coefficient, and the type of solvents [117]. Antioxidant activity of hydrophobic amino acids has been attributed to their ability to interact with lipid molecules by increasing the solubility of peptides in lipids and scavenging lipid-derived radicals through electron donating substituents such as OH and NH₂ on amino acid side chains in peptides. The presence of hydrophobic amino acids can improve the antioxidant activity of peptides by providing a potential pool of free electrons. Additionally, it is believed that aromatic amino acid residues might possess antioxidant properties due to the chelating ability of the imidazole ring and the trapping ability of lipids [116,117].

Peptides containing proline-rich sequences have been identified to possess antioxidant properties. Proline has an electron-rich nitrogen-containing pyrrolidone ring that stabilizes the radical peptide formed after electron donation [118,119]. Peptides with Pro at the C-terminus (e.g., Pox: Tyr-Tyr-His-Pro) were the most potent antioxidant (0.8 TE at 2.5 μM). Modification of the structure by moving Pro into positions X1 (Pro-Tyr-Tyr-His), X2 (Tyr-Pro-Tyr-His), and X3 (Tyr-Tyr-Pro-His), but leaving the other residues in the same order as in Pox, resulted in a significant difference in ORAC which was 0.2, 0.1, and 0.55 TE at 2.5 μM for X1, X2, and X3, respectively [120]. Strong antioxidant activity has been reported for small peptides containing amino acid residues such as Tyr, His, and Pro [121,122]. The dipeptide Tyr-Tyr at the N-terminal position of Tyr-Tyr-His-Pro and Tyr-Tyr-Pro-His was the portion responsible for stronger antioxidant activity. However, Tyr-Pro-Tyr-His showed the weakest ORAC, inhibited ROS production by 36% at 0.07 μM in human keratinocyte cells (HaCat) after treatment with H₂O₂ when compared to Tyr-Tyr-His-Pro (with the highest ORAC) which showed almost similar inhibition of ROS production at 2.5 μM (40%) [120]. It is believed that besides amino acid composition and sequence, the changes of secondary structure have a significant impact on the capture and dissipation of free radicals. The nanopeptide Val-Leu-Leu-Tyr-Lys-Asp-His-Cys-His (1127 Da) produced using self-assembly of pine nut Val-Leu-Leu-Tyr (506 Da) and sea cucumber Lys-Asp-His-Cys-His (638 Da) had significanly higher antioxidant activity compared to individual peptides due to changes in the secondary structure as seen in the lower electron paramagnetic resonance (EPR) signal, higher random crimp degree, and increased supply of hydrogen protons (i.e., the higher exposure to active hydrogen) from Raman spectroscopy and ¹H NMR spectrum analysis in the nanopeptide than the tetrapeptide and the pentapeptide. The DPPH radical scavenging activity at 3 mmol/mL were 6.1, 9.4, and 80.7% of the 4, 5 and 10 amino acid peptides, respectively [123].

The presence of Tyr at the N-terminal position of peptide Tyr-Ala-Glu-Glu-Arg-Tyr-Pro-Ile-Leu has been reported as the residue that most contributed to antioxidant activity (3.8 μM TE/mg protein). However, Tyr-Pro-Ile and Tyr-Gln-Ile-Gly-Leu with Tyr at the same position showed lower ORAC (1.6 and 1.7 μM TE/mg protein, respectively), indicating the role of adjacent amino acids and chain length on antioxidant activity [124]. Tyr-containing peptides from abalone viscera showed strong ABTS radical scavenging activity in the order of Cys-Ile-Gly-Tyr-Asp-Arg (0.144 mg/mL) > Tyr-His-Gly-Phe (0.268 mg/mL) > and Gly-Cys-Tyr-Val-Pro-Lys-Cys (0.389 mg/mL). The first and last peptides, which contained both Tyr and Cys showed similar trends for scavenging DPPH radicals (IC₅₀ of 0.207 and 0.405 mg/mL, respectively). Despite the observed high ABTS radical scavenging activity, peptides Met-Glu-Thr-Tyr and Tyr-His-Gly-Phe which had Tyr at the C- or N-terminal position, respectively, had weak scavenging activity against DPPH radicals (<20%) which was attributed to the lack of Cys in their sequence. Despite having different size or amino acids residues, peptides Gln-Cys-Val-Arg and Gln-Ser-Cys-Ala-Arg-Phe showed similar DPPH radical scavenging activity (IC₅₀ of 0.392 and 0.416 mg/mL, respectively), indicating the complexity of the relation between the peptides’ structures, and function. Regarding the number of amino acid residues within peptide sequence, The peptide Gly-Cys-Tyr-Val-Pro-Lys-Cys, containing two Cys residues showed lower free radical scavenging activity (IC₅₀ of 0.389 and 0.405 mg/mL for scavenging ABTS and DHHP radicals, respectively) than Cys-Ile-Gly-Tyr-Asp-Arg, which contained only one Cys (IC₅₀ of 0.144 and 0.207 mg/mL for scavenging ABTS and DPPH radicals, respectively [125]. Thus, a simple relationship between the number of Cys and antioxidant effects of peptides is not clear and is more related to its position in the peptide chain and the type of amino acids adjacent to it. Generally, the thiol group of Cys have an antioxidant activity that works by donating hydrogen from the SH group or the loss of an electron from its sulfur atom, thus neutralizing free radicals [24].

It was shown that the presence of a Tyr, Trp, Cys, or Met residue with electron/hydrogen donating ability was the driving force for dipeptides to scavenge radicals. The presence of Tyr, Trp, and Cys in the sequence was required for dipeptides to scavenge ABTS^•+, while the presence of Tyr, Trp, Cys, and Met was needed for dipeptides to scavenge ROO^• when using the ORAC assay. Structure-activity relationships showed that Tyr and Trp-containing dipeptides with Tyr/Trp residue at the N-terminus (Tyr/Trp-X; Tyr-Gly, Tyr-Ser, Tyr-Gln, Tyr-Glu) had stronger ORAC and ABTS^•+ scavenging activity than that at the C-terminus (X-Tyr/Trp; Gly-Tyr and Glu-Tyr) and the steric effects, hydrophobicity and hydrogen bonding also affected the neighboring AA. Tyr-containing dipeptides showed higher ABTS^•+ scavenging activity. In contrast, Trp dipeptides (Trp-Gly, Trp-Ser, Trp-Gln, Trp-Glu, Gly-Trp, Glu-Trp) had higher ORAC, and only Cys-containing dipeptides showed moderate reducing power activities [126]. Calculation of BDE, IP, PA, and ETE of Tyr/Trp-X and X-Tyr/Trp (where X was Gly, Leu, Pro, Phe, Ser, Thr, Asn, Gln, Asp, Glu, Lys and Arg) showed that there were little differences among dipeptides, indicating that the neighboring AA did not affect the intrinsic hydrogen or electron-donating ability of the dipeptides studied. Thus, the differences in their radical scavenging activity can be attributed to other factors (such as steric effects, and inter/intra-molecular hydrogen bonds). Furthermore, BDE and PA of Tyr-containing dipeptides were much lower. At the same time, its IP was higher than Trp-X or X-Trp, indicating that HAT and SPLET mechanisms were more favorable for Tyr-containing dipeptides, while SET-PT was the primary mechanism of antioxidant activity of Trp containing dipeptides [126].

Aromatic and acidic amino acids are effective proton and electron donors to neutralize free radicals [127]. The high DPPH^• and OH^• scavenging activity of Glu-Ala-Pro-Val-Glu-Gly-Gly-Leu-Phe-Asp-Tyr-Val-Lys from scallop has been attributed to the presence of the acidic amino acids (Glu and Asp) and two aromatic amino acids (Phe and Tyr) in the sequence. At the same time, Arg at the C-terminal contributed to high ORAC (6.18 μM TE/μmol) of Lys-Leu-Ala-Asp-Met-Leu-Asn-Pro-Glu-Arg [128]. Tyr and Phe in Gly-Glu-Tyr-Gly-Phe-Glu and Phe in Gly-Ile-Glu-Leu-Phe-Pro-Gly-Leu-Pro sturgeon cartilage contributed to higher DPPH^• and OH^• scavenging activity [129].

Antioxidant activity of His-containing peptides has been reported and attributed to the chelating and lipid radical-trapping ability of the imidazole ring [130]. Removal of His from the C-terminal position of Val-Asn-Ala-Val-Leu-His (MW: 651 Da) significantly decreased DPPH (~20%) and ABTS (~16%) radical scavenging activity of the modified peptide Val-Asn-Ala-Val-Leu (MW: 514 Da) compared to ~33% and 21% for the original sequence. From circular dichroism (CD) spectroscopy, the secondary structure of Val-Asn-Ala-Val-Leu had no α-helix with a low band intensity at 195 nm, probably due to shortening of the peptide size after removing the C-terminal His. Treatment of the original and modified peptides with 40 kV/cm pulsed electric field (PEF), transferred β-sheet to the random coil. This led to higher antioxidant activity, and Val-Asn-Ala-Val-Leu-His was much more sensitive to PEF than Val-Asn-Ala-Val-Leu, which exposed the leading active site of the C-terminal His by altering the secondary structure of the peptide [122].

Metal-chelating peptides (MCP) can complex transition metal ions, such as Fe²⁺ and Cu²⁺, involved in ROS production using the Fenton and Haber Weiss reactions. Therefore, peptides with metal chelating ability can act as indirect antioxidants, reducing or inhibiting food oxidation, and increasing food shelf life, while reducing the oxidation products associated with age-related diseases [131]. These peptides can also be used as supplements to provide dietary minerals such as Ca, Zn and Fe with high absorption rates [132,133,134]. His, Lys, Arg, Pro, and Gly were abundant in peptides with metal chelating ability [135]. Glu, Asp, and Gly residues were the major AA in the 26 identified peptides from anchovy stick water hydrolysates. Peptides containing these AA formed complexes with Ca ions more effectively [133]. From tilapia skin gelatin, Gly-Pro-Ala-Gly-Pro-Ala-Gly-Glu-Lys (782 Da), Asp-Gly-Pro-Ser-Gly-Pro-Lys-Gly-Asp-Arg (984 Da), Gly-Leu-Pro-Gly-Pro-Ser-Gly-Glu-Glu-Gly-Lys-Arg (1198 Da) and Asp-Gly-Pro-Ser-Gly-Pro-Lys-Gly-Asp-Arg-Gly-Glu-Thr-Gly-Leu (1441 Da) have been purified from trypsin hydrolysates and showed high Fe²⁺ chelating capacity. Each peptide contained one or more acidic amino acids, i.e., Glu and Asp, or basic amino acids such as Lys and Arg. In addition, three of the four Fe²⁺-chelating peptides contained Ser [132]. Asp, Glu, Gly, and Pro were the primary amino acids in sea cucumber metal-chelating peptides [134].

4. Application of Fish By-Products Protein Hydrolysates to Control Oxidative Deteriorations of Seafood

Oxidation of lipids is often the major cause of quality loss of foods during storage, as seen in the changes in color, texture, flavor, and aroma, which impairs sensory and nutritional properties and the shelf-life of foods [28]. Decomposition of the hydroperoxides formed by pro-oxidative metal ions is a driving factor for lipid oxidation, producing highly reactive alkoxyl lipid radicals and hydroxyl ions. Alkoxyl radicals degrade rapidly to form volatile decomposition products often with off-odors [29]. Furthermore, protein carbonyls can be introduced into proteins using a covalent linkage of lipid carbonyls (e.g., protein-bound malondialdehyde). Protein oxidation leads to functional property changes such as decreased solubility, digestibility, and water-holding capacity [30]. On the other hand, the loss of nutrients and myofibrillar water, and the changes in texture are inevitable during frozen storage [135,136]. The formation of ice crystals, associated with cell membrane rupture and muscle fibers, often lead to protein denaturation and undesirable reactions such as aggregation and decrease of solubility, solute concentration (macromolecular crowding), lipid oxidation, and instability of proteins at the ice-water interphase [137,138]. Protein hydrolysates and peptides may be potential antioxidants to reduce oxidation during food storage, thus extending the shelf life [28]. The antioxidant activity of protein hydrolysates was related to amino acid composition, sequence, size, and the amino acid residues at the C- or N-terminal positions [115]. Enzymatic hydrolysis disrupts the tertiary structure of food proteins, leading to the increase of solvent accessibility of peptides to scavenge free radicals and chelate pro-oxidative metal ions. Protein hydrolysates and peptides have been reported to control food oxidation through various mechanisms including inactivating ROS, scavenging of free radicals, chelation of pro-oxidative metal ions, reducing lipid hydroperoxides, and changes of the physical state of foods. At the same time, peptides control the formation of ice crystals and decrease protein oxidation and denaturation during storage, thus showing bi-functional effects in foods, i.e., antioxidant and cryoprotective activities [25,139]. Table 3 shows the effects of peptides and protein hydrolysates from marine by-product sources on the inhibition/reduction of oxidation of seafood lipids and proteins during processing or storage.

5. Conclusions and Future Challenges Facing By-products Upgrading

Marine by-products have been studied as a source of antioxidant peptides for food, feed and nutraceutical applications. Those studies generally recommended them as potential functional ingredients to enhance health and nutrition. However, the differences in composition of by-products, the type of proteases used and different hydrolysis parameters resulted in various end-products from the same protein. To ensure consistency, the process of upgrading at three levels, i.e., by-products, enzymes, and operating parameters must be optimized for each source of by-products.

Different peptides were produced by different enzymes. However, it has not been determined which peptide is a more potent antioxidant in controlling oxidation in which food system? Different foods, due to inherent composition differences (i.e., different amounts of pro-oxidants, oxidation prone substances and internal antioxidant enzymes) will probably have different reactions with antioxidant peptides with a specific structure during storage. Therefore, the effect of peptides with specific structure or protein hydrolysates produced using a specific condition in different food matrices need be investigated.

Although the structure of peptides was influenced by the specificity of the proteases used, most studies used fresh by-products with acceptable initial quality. It is less know if the same peptides structure and function can be obtained using previously stored by-products and how the quality of proteins in refrigerated or frozen by-products as well as associated chemical reactions during storage will affect the hydrolysis and products structure, function, and stability. This area need to be further investigated, especially in by-products with high lipid and blood contents such as herring and salmonid by-products and to understand which fractions are more oxidized and contribute greater to undesirable biochemical reactions during hydrolysis. Several researchers tried to stabilize by-products before up-grading using antioxidants. There is a price to adding synthetic antioxidants or to maintain the initial quality and safety of agricultural wastes as sources of antioxidative extracts. The practical ability to do this for large quantities of by-products, the space required with energy consumption to create low storage temperature are among the issues that make the valorization more complicated.

Few studies have investigated the relationship between the initial microbial and chemical quality of by-products and the safety of the resulting protein hydrolysates for food applications. Residual antibiotics in protein hydrolysates from intensive fish/shrimp farming, the amount of biogenic amines such as histamine as well as the presence of contaminants (such as cadmium, arsenic, mercury and lead) caused by pollution of the sea or culture water are safety issues that should be considered when selecting by-products.

By-products processing should be done near the fish production and processing centers so the hydrolysis of by-products can be done in the shortest possible time. When the quantity of by-products exceeds the capacity of the plant, they should be stored frozen. Therefore, the impact of frozen storage for varying times needs further study on the structure of peptides (amino acid composition, sequence, and size) and the occurrence of undesirable oxidative deteriorations and biochemical changes that may affect its biological functions. There is limited information about the effect of initial protein quality due to processing and storage on the functional and biological activities of hydrolysates and peptides for food applications. Maintaining the consistency of the raw material properties (in terms of composition, freshness and storage) is likely to be necessary to have a consistent end product from each production batch.

Many studies are underway regarding the use of fish protein hydrolysates in food. Nevertheless, the supply of fresh raw materials with acceptable safety, competitive prices with other commercial ingredients from plant and other sources, and the lack of efficient and standardized techniques to transform fish by-products into marketable forms limits their utilization.

Although the role of protein hydrolysates in maintaining the quality of seafood products has been shown, a standard method for its production from a specific source of marine by-products on a pilot or an industrial scale and its industrial application has not yet been undertaken. Yet, industry prefers to use synthetic preservatives with lower price to maintain the seafood products quality during storage. As an example, in case of shrimp processing, the suggested use of hydrolysates to soak shrimp (as whole or peeled) for 1 h (compared to only a few min in case of sulphate additives) to ensure water holding capacity and protein quality is not practical at an industrial scale for shrimp processors. They cannot wait for such a long time for processing of several tonnes of shrimp that come to a plant each day and should be processed with minimum time.

Protein hydrolysates may also affect the sensory characteristics of food. So how to mask or remove the fish smell using encapsulating methods and the cost of such pretreatments on protein hydrolysates and the market demand for it needs to be addressed.