Calcium- and iron-chelating activity and functional properties of salmon by-product proteolysate

In this study, calcium- and iron-binding proteolysate were generated using salmon by-product. It was shown by the result that under the hydrolysis condition including Neutrase, 45 o C, pH 7, E:S proportion of 70 U/g protein and 8 h of hydrolysis, the proteolysate reached the peaks of CaBC and IBC of 220.8 ± 3.1 mgCa 2+ /g protein and 232.3 ± 4.7 µgFe 2+ /g protein, respectively. In pH from 3–8, after heating at 63 °C for 30 min or 93 °C for 30 s (i) its solubility was over 85% and remained over 65%; (ii) FCs and FSs were 1.4–4.7 and 1.8–12.8 times lower than those of albumin; (iii) EAIs and ESIs were 1.1–2.7 and 4.1–11.1 folds less than those of sodium caseinate, respectively; (iv) WHC and OHC reached 2.00 ± 0.09 mL water/g proteolysate powder and 6.47 ± 0.25 mL oil/g proteolysate powder, respectively. Moreover, the < 1 kDa fraction displayed the highest CaBC and IBC. In conclusion, the proteolysate/peptides could be utilized in either easing micronutrient deciency or preparation of functional peptides.


Introduction
Of all the bivalent metal ions, one of the essential minerals for human body is calcium, taking part in nerve conduction, muscle contraction, mitosis, blood coagulation, and structural support of the skeleton (Liu F-R et al. 2013). Moreover, it also plays an important role in several physiological functions including cell proliferation, responses to hormones and the release of neurotransmitters (Guo L et al. 2014). Rickets, osteoporosis, hypertension, obesity and kidney stone were related to calcium de ciency (Lee SH and Song KB 2009a), all of which could be prevented by calcium-forti ed products or calcium supplements in forms of inorganic calcium, organic calcium, amino acid-calcium complexes or calcium-binding peptides.
However, absorption and bioavailability of calcium could be lowered as during gastrointestinal digestion, inorganic calcium supplements such as calcium carbonate or calcium chloride produce calcium phosphate precipitate, causing intestinal side effects such as atulence and bloating (Bronner F andPansu D 1999, Wang L et al. 2018). As for organic calcium such as calcium lactate and calcium gluconate, in clinical trial, low therapeutic effects were observed due to low bioavailability and low proportion of calcium (Wang L et al. 2018). Because of their high cost, propensity to cause unfavorable color and fat oxidation, amino acid-calcium complexes were gradually replaced (Guo L et al. 2014).
Throughout the changes of pH during digestion, with the capability of remaining a soluble form of calcium, calcium bioavailability could be enhanced utilizing calcium-binding peptides as a potential alternative (Liu F-R et al. 2013). Additionally, there is no denying that the advantages of the absorption of small peptides are little energy consumption, accelerating transport speed, and carriers not being easy to saturate (Sun N et al. 2016). With the ability to prevent the precipitation of calcium during digestion, CPP increases the amount of soluble calcium available for absorption across the mucosa, hence it can be used as a potential calcium supplement (Jung WK and Kim SK 2007). Despite the relatively high price and the complexity of production of CPP, recently, the search for alternative calcium supplements has attracted many researchers. It has been proven that aquatic products and their by-products are great protein sources for the production of calcium-binding peptides, which were found in the proteolysate of Antarctic krill (Hou H et al. 2018), Paci c cod bone (Peng Z et al. 2017), cuticle of the cray sh (Inoue H et al. 2004), tilapia (Charoenphun N et al. 2013) and hoki frame (Jung WK and Kim SK 2007).
Another one of the most essential elements required for energy production, oxygen and electron transport, mitochondrial respiration and DNA synthesis is iron (Hershko C 2005). Iron de ciency could result in anemia, poor cognitive development and increased maternal mortality, all of which could be averted by fortifying iron into food in forms of iron salt, elemental iron and iron-binding peptides (Guo L et al. 2013, Ying L et al. 2017. It was reported from Gaucheron F (2000) that the drawbacks of iron salt and elemental iron that narrowed their applications were low bioavailability, poor taste and insolubility. In addition, during peptic digestion, the ferrous ions must be rst reduced by the enzyme ferric reductase locating on the brush border of the enteric cells before being absorbed as they are quickly oxidized to the insoluble ferric form (Eckert E et al. 2016). Moreover, ferrous sulfate was indicated from the study of Caetano-Silva ME et al. (2018) to promote the formation of hydroxyl radicals, starting the peroxidation of lipids from biologic membranes, enzyme inactivation, and DNA damage. In contrast, although ironbinding peptides were released at higher pH of duodenum, they had the ability to form a stable soluble complex with ferrous ion and remain the complex under acid condition in the stomach, improving iron absorption, stability and bioavailability (Bouhallab S et al. 2002, Lee SH andSong KB 2009b). Recently, several publications have been published on nding iron-binding peptides from various marine life forms and by-products such as shrimp by-product (Huang G et al. 2011b), Alaska pollock skin (Guo L et al. 2013), Paci c cod skin gelatin (Wu W et al. 2017), sea cucumber (Sun N et al. 2017) and anchovy muscle protein (Wu H et al. 2012).
Enzymatic hydrolysis has been an effective strategy in the production of protein hydrolysate possessing functional properties including solubility, heat stability, emulsifying and foaming property, oil and water holding capacity. Proteolysate with size-varied peptides that can be applied in food enhancement could be generated from modi ed enzymatic proteolysis (Pacheco-Aguilar R et al. 2008), which was also considered as a means to convert wastes and by-products into more marketable and acceptable preparations, especially bioactive proteolysate or peptides from sh by-products (Souissi N et al. 2007).
Salmon processing industry has got rid of a colossal amount of by-products containing ns, tail, bone, remaining esh,…, which was evaluated to be abundant in protein, bioactive components and essential nutrients that could improve human health (Chalamaiah M et al. 2012). Nevertheless, as the use of these derivatives is limited, a valuable source of protein becomes a waste. The by-product have been utilized to produce antioxidant peptides (Vo TDL et al. 2018a, Wu R et al. 2017 or proteolysates with some functional features such as emulsifying, foaming, and oil holding capacity (Gbogouri GA et al. 2004, Kristinsson HG andRasco BA 2000). However, publications on salmon derivatives to generate calciumbinding and iron-binding proteolysate expressing functional characteristics have not been reported.
In this investigation, the proteolysate expressing the most remarkable CaBC and IBC was obtained by examining the effects of enzymatic hydrolysis condition comprising of hydrolysis enzyme type, temperature, pH, E:S ratio and hydrolysis time on CaBC and IBC of the salmon by-product proteolysate.
The evaluations of functional properties including solubility, heat stability, emulsifying and foaming property, oil and water holding capacity WHC and OHC of the proteolysate were also carried out. Then four peptide fractions were collected and tested for their CaBC and IBC by performing peptide fractionation of the proteolysate using ultra ltration centrifugal devices.

Materials
Salmon by-product The salmon frames including bones, ns, tails and some remaining esh attached to the frames were procured from a company in Ho Chi Minh city. The by-products were conveyed on ice to the Biochemical laboratory of Ho Chi Minh city University of Technology -Vietnam National University Ho Chi Minh city within 4 h, separately packaged in polyethylene bags, marked and kept at − 20 °C until used. The chemical composition of salmon by-product which contained 61.9 ± 0.2% moisture, 44.3 ± 0.7% crude protein, 45.4 ± 1.1% crude lipid and 10.2 ± 0.2% ash (on dry weight basis) was evaluated by the method of AOAC (2000).

Chemicals and enzyme preparations
Alcalase ® 2.5L, Neutrase ® 0.8L, Protamex ® and Flavourzyme ® 500MG were gained from Novozymes (Denmark) and AB enzymes (Germany). Chemicals were acquired from Sigma-Aldrich and Merck. All reagents were in analytic quality. Double-distilled water was employed in tests.

Preparation of hydrolysate
The procedure of Vo TDL et al. (2018b) with slight modi cation was performed to prepare the hydrolysates. The endogenous enzymes were deactivated by heating mixture of by-product and water with the ratio of 1:10 (w/v) at 90 ºC for 10 min. After hydrolysis at a predetermined time, the enzymes were deactivated by heating the hydrolysates for 10 min at 90 ºC. The supernatant was recovered by centrifuging the hydrolysates. The protein content and DH of the proteolysate were measured applying the method of Lowry OH et al. (1951) and Nielsen PM et al. (2001), respectively. The collected supernatants were freeze -dried using a freeze -dryer (Alpha 1-2/Ldplus, UK) and stocked at − 20 ºC until being used.
Effect of hydrolysis condition on the CaBC, IBC and DH of the proteolysate A single factor test approach, which was carried out by one factor being altered on various levels while others being given, was performed to determine the impacts of ve effective parameters consisting of hydrolysis enzyme type, pH, temperature, E:S ratio and hydrolysis time on the CaBC, IBC and DH of the proteolysate.

Determination of CaBC
The same method in our previous study was employed to perform the calcium-binding test (Vo TDL et al. 2018b). After demineralization using macroporous resin (Amberlite IRC -748I, sodium structure, Acros), hydrolysates with different concentrations up to 500 mg/L were blended with 5 mM CaCl 2 and 20 mM sodium phosphate support (pH 7.8), which was then mixed at 22 o C within 30 minutes. The pH was kept at pH 7.8 utilizing 1N HCl or 1N NaOH. The amount of calcium in supernatants was quanti ed utilizing ortho -cresolphtalein complexone (oCPC). The analysis was performed in triplicate and data were shown as means ± standard deviations.

Determination of IBC
The IBC of the proteolysate in this research was estimated employing the method published in our previous study (Vo TDL et al. 2020). After using macroporous resin (Amberlite IRC-748I sodium form, Acros) for demineralization, 1 mL of proteolysate was added to 2.5 mL of acetate buffer (0.1 M, pH 5) and 0.6 mL of FeSO 4 solution (0.2 mM), in order. After 30 minutes, 0.3 mL of Ferrozine (5 mM) was added and the observation of the absorbance of Fe 2+ -Ferrozine complex was performed at 562 nm. The IBC was calculated by the following formula: Where: A c denotes the absorbance of the blank; A s is the absorbance of the sample; m Fe2+ indicates the initial weight of Fe 2+ , µg; m protein represents the weight of protein of proteolysate, g.

Determination of solubility
The method of Li X et al. (2012) was slightly modi ed to evaluate the solubility of the proteolysate. 1 M HCl or NaOH solution was used to adjust the pH of the mixture of hydrolysate samples (100 mg) in 10 ml of deionized water to 3, 4, 5, 6, 7 and 8. Then, prior to being centrifuged to collect the supernatant, the mixture was stirred at room temperature for 30 min. The sample was next dissolved in 0.5M NaOH solution, of which the total protein content was estimated using the method of Lowry OH et al. (1951). Protein solubility was calculated as follows: Solubility (%) = Protein content in supernatant * 100 / Total protein content in sample (2) Determination of heat stability The heat stability of the proteolysate was assessed utilizing the procedure of Li X et al. (2012) with a minor variation. The pH of the solution consisting of 100 mg of hydrolysate and 10 mL of deionized water was regulated in the range of 3 to 8 using 1 M HCl or 1 M NaOH solution. After being heated at 63 o C for 30 min and 93 o C for 30 s, the supernatant was collected by centrifugation after keeping the solution in iced-water for 10 min. The heat stability of the proteolysate was expressed as its solubility after heat treatment.

Determination of FC and FS
The method of Li X et al. (2012) with a minor variation was performed to estimate the FC of the proteolysate. The FC was assessed in the pH ranging between 3 and 8. The collected whipped specimen by homogenizing 40 mL of 10 mg/mL proteolysate was instantly moved into a 100 mL cylinder and the total volume was recorded after 30 s. The FC was determined by the following equation: (2) Where A indicates the volume of the proteolysate after being whipped (mL); B represents the volume of the proteolysate before being whipped (mL).
The whipped sample was left at 20 o C for 3 min and its volume was then monitored. The FS was evaluated via the following equation: (3) Where A t expresses the volume of the proteolysate after standing (mL); B depicts the volume of the proteolysate before being whipped (mL).

Determination of EAI and ESI
In order to determine emulsifying property of the proteolysate, the method of Li X et al. (2012) with a minor alteration was applied. The mixture of 5 mL of vegetable oil and 15 mL of 10 mg protein/mL hydrolysate had its pH varied from 3 to 8 using 0.1 M NaOH or 0.1 M HCl solution prior to being homogenized. At 0 min and 10 min after homogenization, a 50 µL aliquot of the emulsion was taken from the bottom of the container. Afterwards, 4.95 mL of 1 mg/mL sodium dodecyl sulfate solution was added to the blend and the absorbance was determined at 500 nm. The EAI and ESI were estimated by the equation: EAI (m 2 /g) = 2 * 2.303 * A 0 / (0.25 * protein weight (g)) (4) ESI (min) = A 0 * ∆t / ∆A (5) Where ∆A = A 0 -A 10 ; ∆t = 10 min; A 0 and A 10 are the absorbances of the samples taken at 0 min and 10 min after homogenization, respectively.

Determination of OHC
The procedure of Putra SNKM et al. (2018) was slightly adjusted to assess the OHC of the proteolysate. The mixture of 0.5 g of each sample to 10 mL of vegetable oil in 50 mL centrifugal tube was kept at temperature of 25 ± 1 o C (30 s of agitation was performed every 10 min) for 30 min. Afterwards, it was centrifuged and the volume of the supernatant was measured. The amount of oil clinging to the wall of the tube was evaluated applying the same protocol. This assay was carried out in triplicate and the OHC was described as the volume (mL) of oil absorbed by 1 g of proteolysate.

Determination of WHC
The WHC was determined applying the centrifugation method presented by Putra SNKM et al. (2018). 0.5 g of sample was rst rehydrated with 20 mL of water in each centrifuge tube then scattered with a vortex mixer for 30 s. Next, before centrifugation, the dispersion was left at room temperature for 6 h.
Whatman No. 1 lter paper was used to lter the supernatant and its volume was accurately measured. The variance between the initial volume of distilled water added to the protein sample and the volume of the supernatant was evaluated, which was depicted as mL of water absorbed per 1 g of protein sample.

Statistical analysis
Data were presented as means ± standard deviations of triplicate analyses. Estimation of variance (oneway ANOVA) was run on the data, and the signi cation was a rmed by the Tukey method (p < 0.05). This assessment was ful lled via the Statgraphics Centurion 18 software.

Results And Discussion
Effect of hydrolysis condition on CaBC, IBC and DH of salmon by-product proteolysate Effect of hydrolysis enzyme type It can be shown in Fig. 1A that Neutrase expressed the greatest potential among Flavourzyme, Alcalase and Protamex to gain proteolysate expressing the highest CaBC and IBC of 115.3 ± 1.98 mgCa 2+ /g protein and 123.7 ± 3.5 µgFe 2+ /g protein, respectively. Enzyme speci city plays a vital role in bioactivity of proteolysate as it remarkably impacts molecular size and the hydrophobic/hydrophilic balance of peptides in proteolysate (Kristinsson HG and Rasco BA 2000). Neutrase proteolysate has the highest DH (Fig. 1A), producing more short-chain peptides. This might be due to this protease is an endopeptidase having no speci c cutting site (Yu Y et al. 2018). As for the other three enzymatic proteolysates possessing lower mineral-chelating capability, it might be because of the steric barrier of long chain peptides lowering their migration capacity to chelate target metal ions (Intarasirisawat R et al. 2012).
Hence, Neutrase was implemented for further investigations.
[ Fig. 1 about here. ] Effect of hydrolysis temperature With regards to the impact of temperature, the CaBC and IBC reached the peaks of 157.9 ± 1.3 mgCa 2+ /g protein and 162.6 ± 6.2 µgFe 2+ /g protein, respectively, at 45 o C (Fig. 1B). Temperature affects the activation energy, speed of reaction, and thermal stability of enzyme and substrate, as active site of enzyme is restricted by higher temperature, interrupting hydrolysis and affecting the bioactivity of the proteolysate (Thiansilakul Y et al. 2007). At 40 o C (Fig. 1B), higher DH meant that proteolysate may contain numerous small peptides, which was revealed to be unable to form complexes with metals (Intarasirisawat R et al. 2012). Thus, hydrolysis temperature of 45 o C was chosen for more tests.

Effect of pH
In terms of enzymatic reaction kinetics, charged amino acid residues bind to substrate at the active site of enzyme, forming a speci c association that converts a substrate to a product (Shu G et al. 2017). The calcium-binding capacity of the proteolysate varies as the environmental pH may alter structure of enzyme as well as enzyme-substrate linkage (Shu G et al. 2017). Figure 1C demonstrated that at pH 7 the greatest CaBC of 204.9 ± 2.5 mgCa 2+ /g protein and IBC of 224.9 ± 10.3 µgFe 2+ /g protein were obtained, while at greater or lower pH, proteolysates had lower bioactivity. This was thought to be the changed catalysis activity as a result of the successive deformation of the active site of enzyme (Thiansilakul Y et al. 2007). Therefore, pH 7 was chosen for additional experiments.

Effect of E:S ratio
It was shown in Fig. 1D that the CaBC and IBC of the salmon by-product proteolysate increased correspondingly with the E:S ratio in the E:S ranging from 40 to 70 U/g protein. It could be understood that the higher the E:S proportion is, the higher the proteolysis rate is. As a result, with a higher amount of freed small peptides, the generated proteolysate could achieve a greater DH, improving bioactivity of the proteolysate. However, when the hydrolysis speed is constant, all substrates are converted to products. Therefore, the increase in E:S ratio does not affect the metal-binding capacity (Kang PY et al. 2018). An insigni cant change in DH was shown in Fig. 1D when increasing the E:S ratio from 70 to 80 U/g protein.
Moreover, the bioactivity of proteolysate was assumed by Chen D et al. (2014) to reduce because of the damaged calcium-binding peptides caused by a greater amount of enzyme created through early steps of hydrolysis. So, the E:S ratio of 70 U/g protein was used for further analyses.

Effect of hydrolysis time
As depicted in Fig. 1E, at hydrolysis time of 8 h, CaBC and IBC peaked at 220.8 ± 3.1 mgCa 2+ /g protein and 232.3 ± 4.7 µgFe 2+ /g protein, in order. Enzymatic hydrolysis generates amino acid residues expressing the capability of binding calcium ions (Liu Q et al. 2010), increasing the CaBC of the proteolysate. However, when prolonging the hydrolysis time, the proteolysate's bioactivity is lowered due to either the deeper cleavage of the enzyme on generated peptides or the decrease of the enzyme activity.
Similar results for tilapia protein hydrolysate on the relation between hydrolysis time and the CaBC were observed by Charoenphun N et al. (2013). Similarly, Guo L et al. (2013)  Functional properties of salmon by-product proteolysate Solubility One of the most essential functional features of protein and proteolysate is solubility, which has an impact on other functional properties such as emulsi cation and foaming capability (Putra SNKM et al. 2018). The solubility of proteolysate is enhanced by the hydrogen bonds between water and the smaller peptide fractions, which are formed by cleaving protein using enzymatic hydrolysis, with polar residues (Gbogouri GA et al. 2004). Besides, Kristinsson HG and Rasco BA (2000) reported the enhanced solubility of proteolysate thanks to the balance between hydrophilic and hydrophobic force of peptides. Moreover, the environmental pH affects the solubility of proteolysate by changing the charge of peptides, producing a proteolysate with the lowest solubility at isoelectric point and the greatest solubility when being charged maximally (Kristinsson HG and Rasco BA 2000). Among other functional practices including emulsions, foams and gels, the most important factor is great solubility of protein as soluble proteins assisting homogeneous dispersibility of molecules in colloidal systems, the interfacial features are improved (Ktari N et al. 2012). As can be seen in Fig. 2A, in the pH range from 3 to 8, salmon by-product proteolysate possessed solubility of over 85%, achieving a peak of 96.97 ± 1.51% at pH 8, which was 1.2 times higher than that of Alcalase proteolysates from golden apple snail (Putra SNKM et al. 2018). The dissimilarity in the solvability is explained by the size, the hydrophobic-hydrophilic balance and the charge of peptides formed through enzymatic hydrolysis (Ktari N et al. 2012). The lowest solubility of the protelolyates obtained at pH 4 was thought to be approximate to the isoelectric point of salmon protein.
The result suggested that the salmon by-product proteolysate might be employed in food products for features enhancement.

Heat stability
Heat treatment, as a general unit operation in food processing, can impact their functional features due to thermal sensitiveness of protein (Li X et al. 2012). An e cient index of the denaturation level of protein is protein solubility characterization during heat handling, aiding in controlling emulsi cation, foaming, extraction, and gelation processes (Zayas JF 1997). Heat durability of proteolysate was described as its solubility after thermal processing at a xed condition. In this study, after heat treatment at 63 o C for 30 min (over 85% in the tested pH range), a small change in the solubility of the proteolysate was observed, which was greater than 65% after heat handling at 93 o C for 30 s (Fig. 2A). The heat stability of the proteolysate was relatively lower than that of grass carp proteolysate (Li X et al. 2012). It could be explained by the poorer balance between hydrophilic and hydrophobic force leading to protein aggregation during heat handling (Li X et al. 2012). Additionally, the variance of heat stability of proteolysate is also affected by different protein sources from dissimilar sh species. Heat-stable proteins were reported by Nurdiani R et al. (2016) to be found in shes inhabiting the water with high ambient temperature. Besides, it was published by Nourmohammadi E et al. (2017) that the increase in heat stability may be because of the hindrance to form the secondary structure by Pro, Ileu and amino acids with identically charged side chains as well as the generation of internal hydrophobic bonds in protein molecules.

Foaming property
The foam formation is affected by transportation, penetration as well as rearrangement of molecules at the air-water interface (Intarasirisawat R et al. 2012). Protein molecules must rapidly migrate to, unfold and reorganize at the air-water interface to express great FC (Putra SNKM et al. 2018). In addition, Li X et al. (2012) reported that with the lower solubility of proteolysate, the migration speed of protein molecules to the air/water interface was lower, reducing the FC. Besides, the pH was revealed by Naqash SY and Nazeer RA (2013) to have impacts on the FC of the proteolysate through the net charge of peptides in proteolysate. Figure 2B demonstrated that at pH 3, the FC of the protein hydrolysate reached a peak of 36.31 ± 1.72%, which was 1.9, 2.0 and 1.5 times higher than those of proteolysates from sole skin, squid skin and round scad, respectively . These variances were attributed to the amount variation of longer chain peptides created through enzymatic hydrolysis, which could generate a thicker and stronger lm covering air bubbles (Intarasirisawat R et al. 2012). Furthermore, in the pH ranging from 3 to 8, the FC values of the salmon by-product were 1.4-4.7 times lower than those of albumin solution.
As intermolecular cohesiveness as well as elasticity of the protein polymers are essential for generation of stable foams, the protein molecules are required to create intermolecular polymers that embrace air bubbles to boost the FS (Ktari N et al. 2012). The FS is quanti ed by the level of protein-protein interaction within the matrix correlating with the ionic repulsion of peptides (Naqash SY and Nazeer RA 2013). In this research, at pH 3, the proteolysate exhibited the highest FS of 18.90 ± 0.86% (Fig. 2C), which was comparable to that of pink perch muscle proteolysate (Naqash SY and Nazeer RA 2013). On the other hand, in the pH of 3-8, the salmon by-product exerted FS 1.9-12.8 folds lower than albumin solution did. Altogether, the salmon by-product proteolysate might be utilized in some products to enhance their foaming capacity.

Emulsifying property
The mechanism of emulsi cation of hydrolysate is thought to be the protective membrane, which is formed through the adsorption of peptides on the surface of freshly formed oil droplets through homogenization, block their coalescence (Gbogouri GA et al. 2004). Protein solubility also affects the emulsi cation via rapid migration to and adsorption at the oil-water interface of protein molecules (Kristinsson HG and Rasco BA 2000). As illustrated in Fig. 2D and Fig. 2E, while EAIs of the proteolysate were 1.1-2.7 times lower than those of sodium caseinate, in the pH range 3-8, the proteolysate exhibited the ESI 4.1-11.1 times higher than sodium caseinate did. At pH 8, the EAI and ESI of the proteolysate peaked at 52.46 ± 0.11 m 2 /g and 136.97 ± 3.85 min, respectively, which were 1.4 and 3.8 folds higher than those of round scad proteolysate, in order (Thiansilakul Y et al. 2007). Great ESI was reported by Pacheco-Aguilar R et al. (2008) to be achieved at high pH because the negatively charged peptides at alkaline pH led to the orientation of the peptides at the oil-water interface. Besides, it was reported by Latorres JM et al. (2018) that as the alkaline pH created the repulsion of negative charges of peptides, emulsifying feature of proteolysate was improved, bene ting their better orientation at the oil-water interface. The emulsifying feature of sh proteolysate was also reported to be directly involved in the surface property, molecular size and hydrophobicity of peptides (Santos SDAd et al. 2011). Proteolysates possessing low DH often comprise of larger peptides assisting their emulsifying feature via maintaining good balance between hydrophilic and hydrophobic groups (Kristinsson HG and Rasco BA 2000). Moreover, it was revealed by Putra SNKM et al. (2018) that different amino acid compositions may result in the variation in emulsifying property of different hydrolysates. Furthermore, the emulsifying stability is improved via the combination of greatly elastic protein layers absorbed on the surfaces of oil droplets, which were created by tertiary proteins and their steric effect, generating strong and thick lms around oil droplets (Intarasirisawat R et al. 2012). As a whole, the proteolysate in this study might be considered to be used in some food products to boost their emulsion feature.

OHC and WHC
OHC, is a critical parameter that affects the taste of the product, which is described as the amount of oil directly bound by protein. The physical entrapment of oil is supposed to be the oil-holding mechanism of proteolysate, as the greater the bulk density of protein is, the higher the OHC is (Kristinsson HG and Rasco BA 2000). Besides, OHC of hydrolysate is also under the effect of other factors such as degree of hydrolysis, the surface hydrophobicity of peptides, and enzyme-substrate speci city (Santos SDAd et al. 2011). Proteolysates expressing high DH consisting of a large amount of small peptides with superior hydrophilicity decrease the interaction between peptide and lipid, reducing OHC (Cumby N et al. 2008).
With the capacity to form hydrophobic bonds to lipid, better hydrophobic peptides express higher OHC, increasing durability of protein-lipid complex (Zayas JF 1997). In this test, the OHC of the Neutrase proteolysate achieved 6.47 ± 0.25 mL oil/g proteolysate powder, 2.8 times greater than that of the hydrolysates from Chinese sturgeon (Noman A et al. 2018). This difference was attributed to the variation in hydrophilic polar side chains of peptides in these proteolysates (Noman A et al. 2018). It was suggested from the nding that the proteolysate powder might be employed to delay phase separation as well as improve palatability and taste retention of some food products (Santos SDAd et al. 2011).
WHC describes the protein capacity of absorbing water and maintaining it against gravitational force within a protein matrix. It has an impact on the texture and integrity of food products such as frozen sh llets or meat (Putra SNKM et al. 2018). The WHC of the salmon by-product proteolysate attained 2.00 ± 0.09 mL water/ g proteolysate powder, which was 2.1, 1.7 and 1.1 times lower than those of proteolysates from golden apple snail (Putra SNKM et al. 2018), cuttle sh muscle (Amiri ZR et al. 2016) and tilapia protein (Foh MBK et al. 2011), respectively. As for deciding the WHC of a proteolysate, Cumby N et al. (2008) uncovered that amino acid pro le and peptide size were important parameters. Lower-molecular-weight peptides are often more hydrophilic, therefore they are more effective in keeping water than larger-size peptides (Cumby N et al. 2008). Furthermore, the amount of adsorbed water is signi cantly affected by the rise in concentration of polar groups encompassing carboxyl and amino groups during enzymatic hydrolysis (Putra SNKM et al. 2018). The result concluded that the salmon byproduct proteolysate might be used as a moisture-keeping agent for some food products.
Determination of the CaBC and IBC of peptide fractions [Fig. 3 about here.] It was demonstrated in Fig. 3 that the < 1 kDa peptide fraction possessed the strongest CaBC and IBC of 520.38 ± 25.75 mgCa 2+ /g protein (1.1 times higher than that of casein phosphopeptide), 998.04 ± 30.38 µgFe 2+ /g protein (1.7 times lower than that of Na 2 EDTA), respectively. This study was consistent with the research of Huang G et al. (2011a) and Vo TDL et al. (2018b) on the < 1 kDa peptide fraction exhibiting the highest calcium a nity. The steric obstacle that decreased the capacity to migrate and chelate target metal ions of long chain peptides led to low metal-chelating activity of large peptides (Intarasirisawat R et al. 2012).

Conclusions
In conclusion, the research assessed the bioactivity of the salmon by-product protein hydrolysate considering CaBC and IBC. In addition to a natural mineral-chelating agent for human body, the proteolysate was discovered as an emulsi er or foaming agent for food products. This original revelation pointed out that the value of the by-product was enhanced by proteolysate with a large potential on practical scale, alongside with reducing its risk of environmental pollution. Nevertheless, further clinical studies and organoleptic evaluations on food forti cation using the proteolysate are needed. Declarations BCV analysed the data and drafted the manuscript. LTTT, LVTN, PTN and DTL performed the experiments. MCT processed the image and formatted the manuscript. PTB made ideas. TDLV managed the overall project, supervised the experimental work, edited and reviewed the manuscript. All authors read and approved the nal manuscript.

Funding
This research is funded by Vietnam National University HoChiMinh City (VNU-HCM) under grant number C2017-20-34.

Availability of data and materials
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.
Ethics approval and consent to participate Not applicable.

Consent for publication
Not applicable.