Identification and Active Evaluation of Antioxidant Peptides from Protein Hydrolysates of Skipjack Tuna (Katsuwonus pelamis) Head

For the full use of fish by-products to produce antioxidant peptides, skipjack tuna (Katsuwonus pelamis) heads generated during can processing were defatted and hydrolyzed using the in vitro gastrointestinal (GI) digestion (pepsin–trypsin system) method and six antioxidant peptides (P1 to P6) were purified from the head hydrolysate (KPH) using ultrafiltration and serial chromatography methods. Six isolated peptides (P1 to P6) were identified as Val-Glu-Glu (VEE, P1), Trp-Met-Phe-Asp-Trp (WMFDW, P2), Asp-Ala-Gly-Pro-Tyr-Gly-Pro-Ile (DAGPYGPI, P3), Trp-Met-Gly-Pro-Tyr (WMGPY, P4), Glu-Arg-Gly-Pro-Leu-Gly-Pro-His (ERGPLGPH, P5), and Glu-Met- Gly-Pro-Ala (EMGPA, P6), respectively, using a protein sequencer and electrospray ionization-mass spectrometer. Among skipjack tuna head hydrolysates, fractions, and six isolated peptides (P1 to P6), WMFDW (P2), WMGPY (P4), and EMGPA (P6) showed the highest radical scavenging activities on 2,2-diphenyl-1-picrylhydrazyl (DPPH) (EC50 values of 0.31, 0.33, and 0.46 mg/mL for WMFDW, WMGPY, and EMGPA, respectively), hydroxyl (EC50 values of 0.30, 0.43, and 0.52 mg/mL for WMFDW, WMGPY, and EMGPA, respectively), and superoxide anion (EC50 values of 0.56, 0.38, and 0.71 mg/mL for WMFDW, WMGPY, and EMGPA, respectively). Moreover, WMFDW, WMGPY, and EMGPA showed strong capability in reducing power and lipd peroxidation inhibition in the linoleic acid system. In addition, WMFDW, WMGPY, and EMGPA can retain strong antioxidant activity at temperatures lower than 60 °C and pH values ranged from 5 to 9. The results showed that six isolated peptides (P1 to P6) from skipjack tuna heads, especially WMFDW, WMGPY, and EMGPA, might be applied in health care products acting as powerful antioxidant agents.


Introduction
Bioactive peptides generally range in molecular size from 2 to 20 amino acid residues and show a large number of potential physiologic functions within biological bodies, such as antioxidant, opioid, antihypertensive, immunomodulatory, antithrombotic, and antibacterial activities [1][2][3][4]. The identification, quantitation and activity evaluation of food-derived bioactive peptides are very important for the characterization of its production, biological value and benefits for consumer health [5,6]. In general, bioactive peptides can be produced during food manufacturing either by the spontaneous hydrolysis of food proteins with endogenous digestive enzymes, or by designing

Preparation of the Protein Hydrolysate (KPH) of Kipjack Tuna Heads
Skipjack tuna heads were thawed, cleaned with tap water, minced, and defatted with the previously mentioned method [6]. In brief, the homogenate was mixed with isopropanol at a ratio of 1:5 (w/v) and left to stand at 30 ± 2 • C with continuous agitation, and the isopropanol was changed after 2.0 h. Finally, the solution was centrifuged at 6000 g for 15 min, and the precipitate was dried at 35 ± 2 • C.
The protein hydrolysate (KPH) of skipjack tuna heads was prepared using the in vitro gastrointestinal (GI) digestion method [10]. Briefly, the defatted powders dispersed in distilled water (DW) (pH 1.5, 1%) were firstly hydrolyzed by pepsin with a protease dosage of 1 g pepsin/100 g defatted powder under the conditions of 37.0 ± 2 • C and pH 1.5. Two hours later, the pH of the degraded solution was adjusted to 7.0 using 1.0 M NaOH solution and further hydrolyzed using trypsin with a protease dosage of 1 g trypsin/100 g defatted powder. After 2 h, the temperature of the solution was increased to 95 ± 2 • C and kept for 10 min. Finally, the solution was centrifuged at 9000 g for 10 min at 4 • C and the supernatant (KPH) was lyophilized and stored in a −20 • C freezer.
The concentrations of KPH and its fractions were measured by the method described by Bradford [26] and represented as mg protein/mL. The degree of the hydrolysis (DH) analysis of protein hydrolysates was performed according to the method described by Yang et al. [13].

Separation of APs by Chromatography Methods
The separation process of APs from UF-1 was performed according to the method described by Yang et al. [13]. Briefly, UF-1 (5 mL, 40.0 mg protein/mL) was injected into a pre-equilibrated DEAE-52 cellulose column (2.6 cm × 120 cm) and eluted with 150 mL DW and 0.1, 0.5, and 1.0 M NaCl solution at a flow rate of 1.0 mL/min. Each eluted fraction (5 mL) was collected and monitored at 214 nm. Finally, six fractions (defined as AC-1 to AC-6) were prepared on the peaks of the 214-nm chromatographic curve, and 5 mL AC-4 solution (20.0 mg protein/mL) was added into a Sephadex G-25 column (2.6 cm × 160 cm) and separated with a flow rate of 0.8 mL/min. Three active factions (defined as GC-1, GC-2, and GC-3) were prepared on the chromatographic curve on the 214-nm absorbance of each eluate (3 mL). Finally, GC-3 (20 µL, 10.0 mg protein/mL) was further purified by a High-performance liquid chromatography (HPLC) column of Zorbax, SB C-18 (4.6 mm × 250 mm) on an Agilent 1260 system (Agilent Ltd., Santa Rosa, CA, USA). GC-3 was eluted with ACN solution containing 0.1% TFA at 1.0 mL/min and the ACN concentration increased from 0 to 50% in 25 min in a linear gradient manner. Six APs (P1 to P6) were prepared on the HPLC chromatographic peaks at 214 nm.

Analysis of Amino Acid Sequence and MW
The amino acid sequences of six APs (P1 to P6) were analyzed using a protein sequencer of Applied Biosystems 494 (Applied Biosystems Inc, Foster City, CA, USA). The MWs of six APs (P1 to P6) were determined by a quadrupole time-of-flight mass spectrometer (MS) coupled with an electrospray ionization (ESI) source, respectively.

Antioxidant Activity
The radical scavenging assays of six APs (P1 to P6) were tested on the method described by Yang et al. [13]. The half elimination ratio (EC 50 ) was used to express the radical scavenging capability of six APs and defined as the concentration of a sample that caused a 50% decrease in the initial radical concentration. The assays of lipid peroxidation inhibition and reducing power were determined by the method described by He et al. [17].

Stability Properties
The stability of APs (P2, P4, and P6) was measured according to the previous method [13]. The thermostability of APs was determined using a water bath for 0.5 h and the temperature was set as 20, 40, 60, 80, or 100 • C. The influences of pH treatments at pH values of 3, 5, 7, 9, or 11 were used to evaluate the pH stability of APs (P2, P4, and P6) at 25 • C for 2.5 h. The EC 50 values of APs (P2, P4, and P6) on hydroxyl radical were used as an index and tested using the methods described in Section 2.5.

Statistical Analysis
The data are expressed as the mean ± standard deviation (SD, n = 3). An ANOVA test was used to analyze the differences in data from different groups, and a p-value of less than 0.05 was considered statistically significant.

Preparation of the Protein Hydrolysate (KPH) of Kipjack Tuna Heads
According to the chemical composition analysis, the major components in skipjack tuna heads were moisture, protein, fat, and ash, with contents of 67.1 ± 3.4, 14.6 ± 0.86, 10.2 ± 0.25, and 6.2 ± 0.93 g/100 g, respectively, on the wet weight basis. The data demonstrated that skipjack tuna heads were suitable for the production of bioactive peptides due to their high protein content.
Proteins of skipjack tuna heads were hydrolyzed by the in vitro GI digestion system, and the prepared hydrolysate (KPH) with a DH of 25.76 ± 1.68% can effectively scavenge DPPH radical with an EC 50 value of 5.36 mg protein/mL. The EC 50 value was lower than those of protein hydrolysates from skate cartilage (13.13 mg protein/mL) [3] and bluefin leatherjacket skin (5.23 mg protein/mL) [12], but higher than those of protein hydrolysates from Tergillarca granosa muscle (3.55 mg protein/mL) [13], tilapia skin (3.66 mg/mL) [16], thornback ray (1.98 mg/mL) [27], and salmon pectoral fin (1.63 mg/mL) [28]. Therefore, the results indicated that the in vitro GI digestion model was effective in preparing bioactive peptides of skipjack tuna heads.

Ultrafiltration
KPH was divided by ultrafiltration with MWCO membranes of 3, 5, and 10 kDa, and four fractions, namely, UF-1 (MW < 3 kDa), UF-2 (3-5 kDa), UF-3 (5-10 kDa), and UF-4 (MW > 10 kDa), were prepared, and their scavenging activity on the DPPH radical is shown in Figure 1. The EC 50 value of UF-1 was 3.24 mg protein/mL, which was significantly stronger than those of KPH (5.36 mg protein/mL), and other three fractions (EC 50 value of 4.68, 8.74, and 11.20 mg protein/mL for UF-2, UF-3, and UF-4, respectively) (p < 0.05). KPH consisted of a wide variety of APs with different molecular sizes. Li et al. reported that hydrolysate fractions with a lower MW were more accessible to free radicals and showed high radical scavenging activity [29]. Therefore, UF-1 was rich in low MW APs and selected for the subsequent chromatographic separation. to free radicals and showed high radical scavenging activity [29]. Therefore, UF-1 was rich in low MW APs and selected for the subsequent chromatographic separation.  Figure 2A showed the anion-exchange chromatographic chart of six fractions (AC-1 to AC-6) from UF-1. According to their adsorption capacity with the DEAE-52 cellulose, AC-1, AC-2 and AC-3, AC-4 and AC-5, and AC-6 were gradually eluted out by an eluent of DW, 0.1 NaCl solution, 0.5 NaCl solution, and 1.0 M NaCl solution, respectively.

Anion-Exchange Chromatography
The data in Figure 2B suggested that the EC50 value of AC-4 (2.58 mg protein/mL) was significantly less than those of UF-I (3.24 mg protein/mL) and other five fractions (EC50 values of 10.65, 4.37, 5.69, 3.18, and 7.26 mg protein/mL for AC-1, AC-2, AC-3, AC-5, and AC-6, respectively) (p < 0.05). Amino acid residues in bioactive peptides with the acidic and hydrophobic groups can be easily adsorbed to the anion-exchange resin (Q Sepharose FF, DEAE, DEAE-52 cellulose and XK 260) and separated from other peptides [6,30]. Moreover, amino acid residues with acidic and/or hydrophobic groups, such as Glu, Met, His, Asp and Pro, are recognized as one of the vital factors for the activity of APs [3,31]. Thus, AC-4 isolated should contain APs with high potential activity and was selected for the following experiment.   Figure 2A showed the anion-exchange chromatographic chart of six fractions (AC-1 to AC-6) from UF-1. According to their adsorption capacity with the DEAE-52 cellulose, AC-1, AC-2 and AC-3, AC-4 and AC-5, and AC-6 were gradually eluted out by an eluent of DW, 0.1 NaCl solution, 0.5 NaCl solution, and 1.0 M NaCl solution, respectively. to free radicals and showed high radical scavenging activity [29]. Therefore, UF-1 was rich in low MW APs and selected for the subsequent chromatographic separation.  Figure 2A showed the anion-exchange chromatographic chart of six fractions (AC-1 to AC-6) from UF-1. According to their adsorption capacity with the DEAE-52 cellulose, AC-1, AC-2 and AC-3, AC-4 and AC-5, and AC-6 were gradually eluted out by an eluent of DW, 0.1 NaCl solution, 0.5 NaCl solution, and 1.0 M NaCl solution, respectively.

Anion-Exchange Chromatography
The data in Figure 2B suggested that the EC50 value of AC-4 (2.58 mg protein/mL) was significantly less than those of UF-I (3.24 mg protein/mL) and other five fractions (EC50 values of 10.65, 4.37, 5.69, 3.18, and 7.26 mg protein/mL for AC-1, AC-2, AC-3, AC-5, and AC-6, respectively) (p < 0.05). Amino acid residues in bioactive peptides with the acidic and hydrophobic groups can be easily adsorbed to the anion-exchange resin (Q Sepharose FF, DEAE, DEAE-52 cellulose and XK 260) and separated from other peptides [6,30]. Moreover, amino acid residues with acidic and/or hydrophobic groups, such as Glu, Met, His, Asp and Pro, are recognized as one of the vital factors for the activity of APs [3,31]. Thus, AC-4 isolated should contain APs with high potential activity and was selected for the following experiment.  Figure 3A indicated that AC-4 was separated into three active components (GC-1 to GC-3) by the Sephadex G-25 column. The DPPH radical scavenging assay showed that GC-3 with an EC50 value of 1.65 mg protein/mL had a significantly stronger activity than AC-4 (2.58 mg protein/mL), and the The data in Figure 2B suggested that the EC 50 value of AC-4 (2.58 mg protein/mL) was significantly less than those of UF-I (3.24 mg protein/mL) and other five fractions (EC 50 values of 10.65, 4.37, 5.69, 3.18, and 7.26 mg protein/mL for AC-1, AC-2, AC-3, AC-5, and AC-6, respectively) (p < 0.05). Amino acid residues in bioactive peptides with the acidic and hydrophobic groups can be easily adsorbed to the anion-exchange resin (Q Sepharose FF, DEAE, DEAE-52 cellulose and XK 260) and separated from other peptides [6,30]. Moreover, amino acid residues with acidic and/or hydrophobic groups, such as Glu, Met, His, Asp and Pro, are recognized as one of the vital factors for the activity of APs [3,31]. Thus, AC-4 isolated should contain APs with high potential activity and was selected for the following experiment.  Figure 3A indicated that AC-4 was separated into three active components (GC-1 to GC-3) by the Sephadex G-25 column. The DPPH radical scavenging assay showed that GC-3 with an EC 50 value of 1.65 mg protein/mL had a significantly stronger activity than AC-4 (2.58 mg protein/mL), and the other two active components (EC 50 values of 5.73 and 3.49 mg protein/mL for GC-1 and GC-2, respectively) (p < 0.05). other two active components (EC50 values of 5.73 and 3.49 mg protein/mL for GC-1 and GC-2, respectively) (p < 0.05). Gel filtration has been used for separating bioactive peptides from protein hydrolysates of aquatic byproducts, such as salmon [28], sardinelle [14], cartilaginous fish [32][33][34], blue mussel [31], skipjack tuna [23], miiuy croaker [8], and cuttlefish [15]. The presented results were in agreement with the report by Li et al. and Pan et al. that peptide components with a low-average MW possess stronger antioxidant activities than their larger counterparts [29,35]. Therefore, GC-3 was suitable for the following purification process.

Purification of APs from GC-3 by Reversed-phase HPLC (RP-HPLC)
Finally, GC-3 was purified using a Zorbax C-18 column on an Agilent 1260 HPLC system and the elution profile is shown in

Molecular Structural Analysis
The amino acid sequences of the six isolated APs (P1 to P6) were analyzed by a protein sequencer and identified as Val

Molecular Structural Analysis
The amino acid sequences of the six isolated APs (P1 to P6) were analyzed by a protein sequencer and identified as Val   Figure 5). Those MW data of the six isolated APs (P1 to P6) agreed well with their theoretical masses (Table 1).

Antioxidant Activity
In the experiment, radical scavenging, reducing power and lipid peroxidation inhibiting assays were used to evaluate the activity of six isolated APs (P1 to P6) from the hydrolysate of skipjack tuna heads, and the experimental data were showed in Table 2 and Figures 6, 7 and 8.

Antioxidant Activity
In the experiment, radical scavenging, reducing power and lipid peroxidation inhibiting assays were used to evaluate the activity of six isolated APs (P1 to P6) from the hydrolysate of skipjack tuna heads, and the experimental data were showed in Table 2 and Figures 6-8.

Lipid Peroxidation Inhibition Assay
Using the linoleic acid model system, the lipid peroxidation inhibition abilities of six isolated APs (P1 to P6) were determined and the lower absorbance at 700 nm illustrates higher antioxidant activity [32]. As showed in Figure 8, the absorbance value of the blank control was significantly higher than those of the six isolated APs (P1 to P6) and the positive control of GSH groups, which indicated that six isolated APs (P1 to P6) and GSH can inhibit the peroxidation of linoleic acid during 7 days incubation. In addition, the absorbance values of P2 and P4 groups was lower than those of the other four APs (P1, P3, P5 and P6). The results indicated that P2 and P4 could serve as the primary and secondary antioxidants to inhibit lipid oxidation in food products.  Figure 9A indicated that the scavenging activity of P2, P4 and P6 on hydroxyl radicals (presented as EC50 value) has the same varying tendency. Heat treatments can greatly affect their activity and the EC50 values of P2, P4, or P6 treated at 20 and 40 °C were significantly (p > 0.05) different from those treated at 60, 80, and 100 °C, respectively (p < 0.05). On the other hand, there was no significant difference

Lipid Peroxidation Inhibition Assay
Using the linoleic acid model system, the lipid peroxidation inhibition abilities of six isolated APs (P1 to P6) were determined and the lower absorbance at 700 nm illustrates higher antioxidant activity [32]. As showed in Figure 8, the absorbance value of the blank control was significantly higher than those of the six isolated APs (P1 to P6) and the positive control of GSH groups, which indicated that six isolated APs (P1 to P6) and GSH can inhibit the peroxidation of linoleic acid during 7 days incubation. In addition, the absorbance values of P2 and P4 groups was lower than those of the other four APs (P1, P3, P5 and P6). The results indicated that P2 and P4 could serve as the primary and secondary antioxidants to inhibit lipid oxidation in food products.

Lipid Peroxidation Inhibition Assay
Using the linoleic acid model system, the lipid peroxidation inhibition abilities of six isolated APs (P1 to P6) were determined and the lower absorbance at 700 nm illustrates higher antioxidant activity [32]. As showed in Figure 8, the absorbance value of the blank control was significantly higher than those of the six isolated APs (P1 to P6) and the positive control of GSH groups, which indicated that six isolated APs (P1 to P6) and GSH can inhibit the peroxidation of linoleic acid during 7 days incubation. In addition, the absorbance values of P2 and P4 groups was lower than those of the other four APs (P1, P3, P5 and P6). The results indicated that P2 and P4 could serve as the primary and secondary antioxidants to inhibit lipid oxidation in food products.  Figure 9A indicated that the scavenging activity of P2, P4 and P6 on hydroxyl radicals (presented as EC50 value) has the same varying tendency. Heat treatments can greatly affect their activity and the EC50 values of P2, P4, or P6 treated at 20 and 40 °C were significantly (p > 0.05) different from those treated at 60, 80, and 100 °C, respectively (p < 0.05). On the other hand, there was no significant difference  Figure 9A indicated that the scavenging activity of P2, P4 and P6 on hydroxyl radicals (presented as EC 50 value) has the same varying tendency. Heat treatments can greatly affect their activity and the EC 50 values of P2, P4, or P6 treated at 20 and 40 • C were significantly (p > 0.05) different from those treated at 60, 80, and 100 • C, respectively (p < 0.05). On the other hand, there was no significant difference in their EC 50 values on hydroxyl radicals when P2, P4, or P6 were disposed at 20 and 40 • C for 0.5 h, respectively (p > 0.05).

Effects of Thermal and pH Treatments on the Stability of P2, P4 and P6
Antioxidants 2019, 8, x 11 of 16 in their EC50 values on hydroxyl radicals when P2, P4, or P6 were disposed at 20 and 40 °C for 0.5 h, respectively (p > 0.05). Figure 9B shows the influences of acid and alkali treatments on P2, P4 and P6. The results indicated that the EC50 values of P2, P4, or P6 showed no significant difference when the pH value ranged from 5 to 9, but they were significantly different from the EC50 values of P2, P4, or P6 treated at pH 3 or pH 11. In addition, the alkaline treatment had a stronger effect on P2 and P6 compared with P4.

Discussion
Proteins in seafoods and their by-products have a high structural diversity and represent a massive resource for the mining of bioactive peptides [1]. Bioactive peptides are released from the sequence of proteins by different hydrolysis methods, such as solvent extraction, microbial fermentation, enzymatic hydrolysis and chemical treatment [5]. Nevertheless, the proteolytic hydrolysis process is the popular method in health care food industries to avoid toxic chemicals or residual organic solvents in the final products. In addition, the specificity and parameters of the proteases significantly affect the MW and amino acid composition of bioactive peptides [47][48][49]. In some cases, specific enzyme combinations are used to enhance the DH of protein and prepare hydrolysates enriched with low MW APs [44,48,50].
The structural activity specialties can provide guides to speculate the activity of isolated APs. The amino acid composition and sequence, MW or molecular size, and spatial structures are deemed to have a vital role in the activity of APs. In general, APs with a low MW have more opportunities to interact with target molecules to terminate free radical reactions and prevent unsaturated lipid peroxidation [44]. The conclusion was confirmed by the literature that the antioxidant capacities of protein hydrolysates or APs were inversely related to the logarithm of their average MW [29,35]. In addition, small APs have a high likelihood to exploit new drugs because they easily pass through the blood-brain barrier to cause effects in human bodies [32]. The six isolated APs (P1 to P6) are  Figure 9B shows the influences of acid and alkali treatments on P2, P4 and P6. The results indicated that the EC 50 values of P2, P4, or P6 showed no significant difference when the pH value ranged from 5 to 9, but they were significantly different from the EC 50 values of P2, P4, or P6 treated at pH 3 or pH 11. In addition, the alkaline treatment had a stronger effect on P2 and P6 compared with P4.

Discussion
Proteins in seafoods and their by-products have a high structural diversity and represent a massive resource for the mining of bioactive peptides [1]. Bioactive peptides are released from the sequence of proteins by different hydrolysis methods, such as solvent extraction, microbial fermentation, enzymatic hydrolysis and chemical treatment [5]. Nevertheless, the proteolytic hydrolysis process is the popular method in health care food industries to avoid toxic chemicals or residual organic solvents in the final products. In addition, the specificity and parameters of the proteases significantly affect the MW and amino acid composition of bioactive peptides [47][48][49]. In some cases, specific enzyme combinations are used to enhance the DH of protein and prepare hydrolysates enriched with low MW APs [44,48,50].
The structural activity specialties can provide guides to speculate the activity of isolated APs. The amino acid composition and sequence, MW or molecular size, and spatial structures are deemed to have a vital role in the activity of APs. In general, APs with a low MW have more opportunities to interact with target molecules to terminate free radical reactions and prevent unsaturated lipid peroxidation [44]. The conclusion was confirmed by the literature that the antioxidant capacities of protein hydrolysates or APs were inversely related to the logarithm of their average MW [29,35]. In addition, small APs have a high likelihood to exploit new drugs because they easily pass through the blood-brain barrier to cause effects in human bodies [32]. The six isolated APs (P1 to P6) are tripeptides to octapeptides with MWs varying from 375.39 to 861.95 Da (Table 1), which shows that they have a high chance of interacting with ROS to prevent lipid peroxidation.
Previous literature indicated that the kinds of amino acids in the APs were regarded as key factors for their activity [5]. The hydrophobic groups of hydrophobic amino acid residues such as Pro, Met, Ala, Leu, and Ile, can highly react with hydrophobic polyunsaturated fatty acids (PUFAs) to inhibit lipid peroxidation in lipid-rich foods [6,8]. Wu et al. confirmed that the inhibiting ability of Pro-Met-Arg-Gly-Gly-Gly-Gly-Tyr-His-Tyr (PMRGGGGYHY) on free-radical chain reactions was attribute to Met residue because it could serve as a reactive site for formatting a sulfoxide structure to scavenge oxidants [51]. The pyrrolidine ring of the Pro residue can increase the antioxidant ability of Pro-His-His (PHH), Leu-Asp-Glu-Pro-Asp-Pro-Leu (LDEPDPL), Gly-Pro-Gly-Gly-Phe-Ile (GPGGFI), and Phe-Ile-Gly-Pro (FIGP), and to scavenge ROS due to its low ionization potential [52,53]. Zhao et al. [8] and Yang et al. [44] reported that the Ala residues contribute to the ability of Gly-Ile-Glu-Trp-Ala (GIEWA) and Ala-Lys-Gly (AKG) on radical-scavenging and lipid peroxidation inhibition. Moreover, the aromatic groups of Trp, Phe, and Tyr residues can provide protons to stabilize electron-deficient radicals [54,55]. Guo et al. reported that the sequence of APs with Trp residues (Trp-Asn-Glu-His and His-Glu-Trp) and Tyr residues (Arg-Tyr, Lys-Tyr, Tyr-Asp, Tyr-Tyr, Arg-Tyr-Asn, Tyr-Asp-Tyr, Tyr-Glu-Gly, and Tyr-Glu-Glu-Asn) showed high antioxidative ability [54]. The Phe residue in KHNRGDEF can serve as a radical scavenger through providing protons to quench unpaired radicals [55]. In consequence, hydrophobic/aromatic amino acid residues in WMFDW (Trp, Met, and Phe), WMGPY (Trp, Met, and Pro), and EMGPA (Met, Pro and Ala) are vital factors for their activity.
The carboxyl and amino groups in polar amino acid residues are important for the hydroxyl radical scavenging and metal ion chelating capacities of APs [8,30]. Acidic amino acid residues of Asp and Glu were reported to have great influence on the antioxidant activities of Asn-Tyr-Asp-Gly-Ser-Thr-Asp-Tyr-Gly-Ile-Leu-Gln-Ile-Asn-Ser-Arg and Leu-Asp-Glu-Pro-Asp-Pro-Leu-Ile [56,57]. In addition, Gly residue can keep the high flexibleness of the polypeptide skeleton and its single hydrogen atom can be donated to neutralize ROS [5,31,58]. Therefore, polar amino acids, including the Asp residue in WMFDW, Gly residue in WMGPY, and Gly and Glu residues in EMGPA, should play a critical role in their scavenging activities of hydroxyl radicals.
Heating and acid-base treatments are common processing methods in the food industry. APs with heat-resistant properties are beneficial to increase the quality guarantee period of food products when they can retain all or major activity after heating treatment, and APs can be used in more kinds of liquid products if they keep their bioactivity in a wide range of pH values [19]. Therefore, the thermal and pH stability of APs are key properties for designing their specific parameters for application in food products [13,36,59]. Wong et al. prepared two antioxidant hexapeptides (Trp-Ala-Phe-Ala-Pro-Ala and Met-Tyr-Pro-Gly-Leu-Ala) from the hydrolysate of blue-spotted stingray and confirmed that the two hexapeptides had excellent stability because no significant difference was found when they were treated at 25-100 • C or at pH values of 3-11 (p > 0.05) [19]. In contrast, Ala-Thr-Ser-His-His from the hydrolysate of sandfish incubated at 50-90 • C reduced its partial DPPH radical scavenging activity. In addition, ATSHH absorbed moderate losses of activity when it was treated at pH 10-12 or pH 2. Yang et al. released homologous conclusions that Met-Asp-Leu-Phe-Thr-Glu and Trp-Pro-Pro-Asp from protein hydrolysates of Tergillarca granosa were not appropriate for high-temperature processing (>80 • C) and alkaline conditions (pH > 9.0) in food products [13]. The present results indicate that P2 (WMFDW), P4 (WMGPY), and P6 (EMGPA) have similar thermal and pH stability with ATSHH, MDLFTE and WPPD, because they can significantly decrease antioxidant activity when they are treated under high temperature (80, and 100 • C) and acid (pH 3) and alkali (pH 9) conditions.

Conclusions
In this experiment, skipjack tuna (K. pelamis) heads were hydrolyzed by in vitro GI digestion and six APs separated from the hydrolysate (KPH) were identified as VEE, WMFDW, DAGPYGPI, WMGPY, ERGPLGPH, and EMGPA, respectively. The six isolated APs, especially MDLFTE, WPPD, and EMGPA, exhibited high radical scavenging, reducing power, and lipid peroxidation inhibition capabilities. In addition, MDLFTE, WPPD, and EMGPA are unstable and can keep effective antioxidant activity when environmental temperatures are lower than 60 • C and pH values range from 5 to 9. The investigation showed that the hydrolysates and APs from skipjack tuna (K. pelamis) heads can be served as antioxidant functional ingredients in health care products.