Collagen Peptides-Minerals Complexes from the Bovine Bone by-Product to Prevent Lipids Peroxidation in Meat and Butter and to Quench Free Radicals—Influence of Proteases and of Steam Sterilisation

Abstract

Food-grade animal by-products are a source of bioactive peptides that could be used as natural antioxidant compounds. Until now, only few studies have put attention to the research of the most effective enzyme(s), to the antioxidant power of collagen peptides in food matrix and to the consideration of a mineralized collagen tissue such as bone, in particular the bovine one. Hence, this study aimed to investigate the effect of enzymes of different origins (animal, plant and microbial) and the effect of processing parameters such as the enzyme-collagen ratio, the steam sterilisation, the concentration of peptides and the mixing of peptides obtained by different enzymes on their antioxidant activity. Low molecular weight peptides (<3000 Da) were produced by the hydrolysis of bovine bone collagen by bromelain, cathepsin B and collagenase B. The antioxidant activity towards lipids peroxidation in meat and butter and the quenching capacity toward free radical were assessed. The bone minerals calcium, magnesium and phosphorous were also present in solution. Peptides-minerals complexes showed a significant antioxidant activity, which depended on the enzyme and on the test used to measure the antioxidant power; collagenase B showed the highest antioxidant capacity, which was further improved by peptides mixing and concentration; sterilization had no adverse effect on the antioxidant capacity. The results of this study can corroborate that food-grade animal by-products, such as the bovine bone, could be valorised for high-value food and health application, which can contribute to a circular use of the agro-food resources.

1. Introduction

Antioxidants and antiradicals compounds can preserve the quality and can contribute to the stability of foods, drugs and chemical products towards oxidation reactions, which can occur during exposure to air (oxygen), heat or light, and during their production, storage and transportation. As such, antioxidant and antiradical molecules continue to be demanded in many fields [1]. In body cells, the oxidative stress has been related to the appearance of diseases such as diabetes, hypertensions, obesity, cardiovascular issues, dyslipidaemia, etc. [1,2], hence, the incorporation of antioxidant compounds in food can contribute to increase its healthy value apart to preserve the nutritional one. Until now, synthetic antioxidants have been more utilized than the natural analogues because they often show higher stability and performance, wide availability and lower costs. However, products of natural origin are increasingly demanded either by consumers or by industry, which are concerned about the presence of synthetic compounds in food [3] and also because some synthetic antioxidants have been associated to health issues such as allergies, gastrointestinal problems, premature senescence, adverse effects on liver and increased risk of cancer [4].

Collagen peptides from animal and marine tissues, such as skin, scales, cartilages, bone, etc., have received attention since they showed a significant potential in multiples fields such as food, nutrition, health, cosmetics and biomaterials [5,6,7,8,9,10,11]. A significant range of bioactivities has been observed, such as antihypertensive, antioxidant, antidiabetic, immunomodulatory, anti-inflammatory, opioid agonist, wound healing, mineral binding, drug carrier, contribution to the synthesis of native collagen, etc. [5,6,7,8,9,10,11]. Most used enzymes for the hydrolysis of collagen to bioactive peptides (<3000 Da molecular weight) have been the digestive enzymes pepsin and trypsin, the specific collagen enzyme collagenase, and alcalase since and it has showed to be effective in the release of bioactive peptides from different protein sources and it is easily available at a commercial scale [7,8,9,11,12]. However, few studies have focused on other enzymes and on the effect on bioactivity of processing parameters such as the enzyme: collagen ratio, sterilisation, heating, and also possible synergies between peptides obtained by several enzymes. Moreover, the antioxidant activity has mainly been addressed by the use of synthetic radicals and very few studies are available on the antioxidant effect on food matrix [13,14]. Besides, among all animal tissues, the mineralised ones, such as bone (and fish scales), have been considered at a lower extent than the non-mineralised tissues (such as skin, tendons and cartilages) and in particular if of bovine origin [7,14]. The animal and fish bone are composed by 60% w/w of mineral matter, mainly calcium and phosphorous and in a lower amount magnesium, by 30% w/w of collagen, mainly of type I, and by 10% of water. In particular, the bone stores 99% of calcium, 85% of phosphorous and almost 60% of all the magnesium of the animal and human body [6]. Animal bone is one of the most abundant by-products, in terms of volume and weight, of meat and milk production. The European regulation allows the valorisation of food-grade bones in several fields, including food, medicine and agriculture (article 33 of the EC 1069/2009, and EC 2019/1009). Nonetheless, for food applications, bones continue to be mainly converted into gelatine albeit at a low degree (ca. 20%) due to presence of the mineral phase; although gelatine, from several animal tissues, has showed a significant antioxidant power, it is not used for this purpose since it significantly affects the texture of the products where it is incorporated [5] besides, it has a moderate commercial value. Bones are currently also applied for pet-food and, depending on the country, even larger quantities are incinerated or exported to limit the disposal costs [15,16]. This is mainly because no high-value valorisation itineraries have been considered yet at a scientific level and neither implemented at an industrial scale. Hence, this study aimed to go beyond the state of the art by assessing the effect of the aforementioned parameters on the antioxidant activity of bovine bone collagen peptides-minerals complexes towards lipids peroxidation in food matrix, such as bovine meat and butter, and the antiradical activity towards synthetic free radicals that are representative of naturally occurring radicals. Eventually, this study also aimed to assess the potential of food-grade animal bones to promote an improved valorisation and for application in fields currently demanding for local and natural sources in a context of a circular bioeconomy.

2. Materials and Methods

2.1. Collagen Hydrolysis

To produce the peptides, collagen was first extracted from the bovine cortical bone (femur of Prim’Holstein cow), and in particular from the femur, collected at the experimental slaughterhouse of INRAE (Saint-Genès-Champanelle, France). The extraction was carried out as described in Ferraro et al. [17], and it allowed for the concomitant extraction of bone minerals, calcium, magnesium and phosphorous. Composition of the bone (femur) extract was as reported in Aubry et al. [13]. Three different enzymes were used to hydrolyse the bone extract: collagenase B (a metalloprotease, of microbial origin and namely from Clostridium histolyticum), cathepsin B (a cysteine protease, from animal origin and namely from bovine spleen) and bromelain (a cysteine protease, of plant origin and namely pineapple stem) [18] (Sigma-Aldrich, Saint-Quentin-Fallavier, France). In

Table 1. Molecular weight, specificity, cutting sites, and optimal pH and Temperature of hydrolysis for the enzymes tested to obtain bovine bone collagen peptides (AA = amino acid) [18].

Enzyme	MW (kDa)	Specificity/Cutting Sites	T (°C)	pH
Collagenase B	68–130	specific/Gly-AA	25	6–8
Cathepsin B	38	non-specific/large hydrophobic AA	37	5
Bromelain	20–30	non-specific/hydrophilic AA	37	7

are reported some characteristic data of those enzymes and the conditions used for the hydrolysis. Enzyme to collagen ratio (E:C) studied was 1/20 and 1/50 w/w. The hydrolysis duration was set at 24 h, after which the enzyme was inactivated through boiling at 95 °C for 5 min. The solutions were then filtered through 3000 Da molecular weight cut-off (Merk Millipore, Molsheim, France) and the fractions obtained were tested for the antioxidant and antiradicals activity. Hydrolysis were done in triplicate.

2.2. Collagen Peptides Quantification

To measure the concentration of collagen peptides, the Bicinchoninic Acid (BCA) test [19] was used; for the purpose, the BCA assay kit developed by Sigma-Aldrich was employed. As the BCA test relies on the reaction between copper (Cu²⁺) and the peptide bond, it can be applied for the determination of either proteins or peptides, starting from a di-peptide [19]. The test was carried out as described in Aubry et al. [13].

2.3. Analysis of Calcium and Magnesium and Estimation of Phosphorous Content

Concentration of mineral cations calcium and magnesium in <3000 Da and >3000 Da fractions, was determined by ion chromatography with the equipment Metrohm 850 Professional IC (Metrohm; Villebon-sur-Yvette, France). The ion exchange resin column Metrosep C4 150 × 4 mm (Metrohm) was used as a stationary phase while the mobile phase was 1.7 mM nitric acid and 0.7 mM dipicolinic acid (Sigma-Aldrich) in ultra-pure water. Detection was done by conductivity. All the determinations were done in triplicate. Phosphorous content was estimated by FTIR (Fourier Transform Infra-Red) spectroscopy and from calcium content; in the bone calcium is present in the form of nanocrystals of calcium phosphates, namely tri-calcium phosphate, Ca₃(PO₄)₂, and, in a lower amount, di-calcium phosphate, CaHPO₄ [20]. Phosphates, in calcium phosphate complexes, show infra-red absorption bands in the ca. 900–1150 cm⁻¹ frequency region, where the PO₄³⁻ group appears at the higher frequencies, 1000–1150 cm⁻¹, and the HPO₄²⁻ group shows a pick at lower frequency values, ca. 900–960 cm⁻¹ [21]. The FTIR equipment TENSOR II (Bruker; Lyon, France) and the software OPUS v.8 were used for spectra acquisition and treatment (noise and baseline correction, and integration).

2.4. Antioxidant Activity of <3000 Da Fractions

Antioxidant activity against lipids peroxidation in food matrix, namely meat and butter, was assessed through the TBArs (Thiobarbituric Acid reactive substances) test. Fresh semimembranosus muscle from Charolaise heifer (5% fat content on raw weight) and cow milk butter (80% fat content) were obtained from supermarket in Clermont-Ferrand (France).

The free radicals quenching activity was evaluated by four different tests, namely ABTS (2,2′-azinobis-3-ethyl-benzothiazoline-6-sulfonic-acid), DPPH (2,2′-Dyphenil-1-picryl-hyrazyl), FRAP (Ferric Reducing Antioxidant Power) and ORAC (Oxygen Radical Absorbance Capacity), which are based on cell-free synthetic radical systems that represent reactive oxygen species such as peroxyl radicals (ROO•), alkoxyl radicals (RO•), superoxide anion (O₂^─), hydrogen peroxide (H₂O₂) and hydroxyl radical (HO•), and nitrogen radicals (N•) [22]. Three factors were considered for the evaluation of the antioxidant activity: (i) the effect of steam sterilisation by comparing the results with the non-sterilised fractions. Sterilisation was carried out by autoclaving at 121 °C, 15 psi, 30 min, with the equipment LS-B75L-II (Socimed; Sains, France); (ii) the effect of a 3-folds concentration of the fractions for ABTS, DPPH and FRAP and the effect of sample volumes for ORAC and TBArs; concentration was attained by speed-vacuum, with the equipment SPD120P1, ThermoFisher (Illkirch-Graffenstaden, France); (iii) the effect of fractions mixing, in equal volumes, as follows: (a) 50% of < 3000 Da fraction obtained by collagenase—50% by cathepsin; (b) 50% by collagenase—50% by bromelain; (c) 50% by bromelain—50% by cathepsin; (d) 1/3 by collagenase—1/3 by cathepsin—1/3 by bromelain.

2.4.1. Thiobarbituric Acid Reactive Substances (TBArs) Test

TBArs test to assess lipids peroxidation in meat and butter was performed according to the method developed by Lynch and Frei [23]. For the purpose, 1 g of raw meat muscle was finely grinded in 9 mL phosphate buffer (pH 7.4) and in ice bath to avoid oxidation. Same quantities were used for butter. For the blank, TBArs values in meat and butter were determined without the addition of the <3000 Da fraction. For the purpose, 500 µL of grinded meat solution (same for butter), were incubated with 500 µL of water, 1% (w/v) 2-thiobarbituric acid in 50 mM NaOH (0.25 mL) and 2.8% (w/v) trichloroacetic acid (0.25 mL) in a boiling water bath for 10 min. After cooling to room temperature, the pink chromogen was extracted into n-butanol (2 mL) and its absorbance at 535 nm was measured. All reagents were from Sigma-Aldrich. To test the effect of sample and of sample volume, 500 µL of grinded meat solution (same for butter) were incubated with 500, 250, 100, 50 and 25 µL of each <3000 Da peptides-minerals fraction. The assay was carried out at ambient temperature with 30 min incubation time, and at 70 °C with several incubation times: 5, 15, 30, 45, 60, 90 and 120 min. TBArs concentration were calculated as equivalents of malondialdehyde (MDA) (mg MDA/Kg of meat of butter), the most prominent product of lipids peroxidation, though its molar extinction coefficient which is 1.56 × 10⁵ M⁻¹ · cm⁻¹. Percentage of reduction of TBArs obtained by the addition of <3000 Da fraction of the bone extract was then calculated with respect to TBArs values without the sample.

2.4.2. ABTS, DPPH, FRAP and ORAC Essays

ABTS, DPPH and FRAP tests were carried out as described in Aubry et al. [13], and according to the methods reported in Guimarães et al. [24], Brand-Williams et al. [25] and Benzie and Strain [26], respectively.

ORAC test was performed by the method described by Ou et al. [27], adapted by Dávalos et al. [28]. The spectrofluorometer Jasco FP 83000 (Jasco; Lisses, France) equipped with a 96-wells microplate reader was used for fluorescence intensity measurements. For the assay, a solution of the standard oxidant AAPH (2,2′-azo-bis-2-methylpropionamide-dihydrochloride) (Sigma-Aldrich), a peroxyl radical, was prepared daily by dissolving 0.13 g in 10 mL sodium phosphate buffer pH 7.4 (SPB) to obtain a concentration of 46.6 mM. The fluorescence tracer compound, fluorescein disodium (FL) (Sigma-Aldrich), was dissolved in SPB to obtain a stock solution at a concentration of 1.17 mM. The working solution of 117 nM FL was prepared daily by dissolution the stock in SPB. FL was used as a probe, where the loss of fluorescence induced by the peroxyl radical AAPH• is a measure of the extent of oxidation. A stock solution of 0.1 mM of the standard antioxidant Trolox was prepared by dissolving 0.0125 g in 1 mL methanol and 49 mL SPB; this stock was then diluted with SPB to obtain the calibration solutions at 0.2, 0.4, 0.6, 0.8, 1, 1.2, 1.4 and 1.6 nM Trolox. The final assay mixtures of 200 μL contained: 80 μL SPB and 120 μL FL for the control; 20 μL SPB, 120 μL FL and 60 μL AAPH for the blank; 20 μL of Trolox calibration solution (at different concentrations), 120 μL FL and 60 μL of AAPH for calibration; 20 μL of each sample mixture (at six different dilutions), 120 μL FL and 60 μL of AAPH for sample. The six sample dilutions were obtained by diluting 20, 50, 60, 80, 100 and 200-fold the <3000 Da fraction with SPB. The fluorescein intensity—485 nm in excitation and 520 nm in emission—was measured at 40 °C, at each minute, over 1 h and 40 min. The relative ORAC value was expressed as μmol TE/L and was determined as follows:

$$\mathrm{ORAC}=\left(\frac{{ \mathrm{AUC}}_{\mathrm{Sample}}-{ \mathrm{AUC}}_{\mathrm{Blank} }}{{\mathrm{mL}}_{\mathrm{Sample}}}\right)\times \left(\frac{{\mathsf{\mu }\mathrm{mol}}_{\mathrm{Trolox}}}{{\mathrm{AUC}}_{\mathrm{Trolox}}-{\mathrm{AUC}}_{\mathrm{Blank}}}\right)\times 1000$$

(1)

where AUC is the area under the curve, the first term represents the slope of the experimental data line and the second term represents the reverse of the slope of the calibration line of Trolox. Assays were done in triplicate. A summary of the tests carried out is reported in

Table 2. Summary of the tests carried out (E:C = enzyme:collagen ratio).

Enzyme	Test	E:C Ratio (w/w)	Incubation Time (min)	T (°C)	Treatment (for Each Test)
Bromelain Cathepsin B Collagenase B	ABTS	1/20, 1/50	30, 60, 90, 120, 180	Amb.	(a) No treatment (b) Sterilisation (c) Concentration/Volume (d) Complexes mixing
	DPPH	1/20, 1/50	30, 60, 90, 120, 180	Amb.
	FRAP	1/20, 1/50	30, 60, 90, 120, 180	Amb.
	ORAC	1/20, 1/50	260 *	37 *
	TBArs	1/20, 1/50	5, 15, 30, 45, 60, 90, 120	Amb., 70

(*) according to the method [22,23].

.

2.5. Free Amino Acids Analysis

Free amino acids content in the <3000 Da fractions was determined by fast high-performance liquid chromatography, with the equipment Acquity Arc, Waters (Saint-Quentin-en-Yvelines, France) coupled with a fluorescence detector (model FLR 2475: ex/em 266/473, Waters). Samples were derivatized using the Waters AccQ-Tag^TM ultra reagent kit. For the purpose, 10 µL of each sample were mixed with a borate buffer containing the internal standard (ABBA^TM). Then, 20 µL of the derivatizing reagent were added to the previous solution and this final mix was heated at 55 °C for 10 min. For the analysis, 2 µL of sample were injected on the column Xbridge C18 (150 × 3.0 mm, 3.5 µm) and eluted at the flow rate of 1 mL/min. The separation gradient was generated using four mobile phases: (A) AccQ-Tag eluant A, (B) water/acetonitrile 90/10 + 2% formic acid, (C) water and (D) acetonitrile +2% formic acid. The gradient was defined as follows: from 0 to 12 min A10%/B0%/C90%/D0%; from 12,1 to 15.5 min A10%/B80%/C10%/D0%; from 15.6 to 16.8 min A8%/B16%/C58%/D18%; from 16.9 to 17.5 min A8%/B0%/C71%/D21%; from 17.6 to 19 min A4%/B0%/C36%/D60% and from 19.1 min to the final time of 22.5 min A2%/B0%/C98%/D0%. Data acquisition and chromatograms treatment were performed with the Waters’ Empower v.3. Amino acids concentration was determined based on amino acids calibration standard run at 8 concentrations: 1.25, 2.5, 12.5, 50, 125, 250, 625 and 1250 µM of Ala, Arg, Asn, Asp, Gln, Glu, Gly, His, Hyl, Hyp, Ile, Leu, Lys, Met, Phe, Pro, Ser, Thr, Trp, Tyr, Val. The proportional molar concentration for each amino acid in the peptides solutions was calculated based on the concentration of the standard amino acids.

2.6. Statistical Analysis

For each treatment, analysis of variance (ANOVA) was performed to assess differences according to the enzyme and to the E:C ratio used. Agglomerative Hierarchical Clustering (AHC) based on the enzyme and on the E:C ratio used, and correlation tests (Pearson) between variables were also performed. Confidence level was set at 95%. The software STATISTICA v.13.3 (Tibco Software; Palo Alto, CA, USA) was used.

3. Results

Bromelain, collagenase and cathepsin had a significantly different effect on the concentration of peptides and of free amino acids in the <3000 Da fraction; the effect of the enzyme/collagen ratio (E:C) was significant too (Figure 1,

Table 3. Free amino acids composition (mg/L) for the <3000 Da peptides-minerals complexes fractions obtained by bromelain (Bro), cathepsin B (Cat) and collagenase B (Col) (values in brackets express concentration in µM).

	E:C 1/20			E:C 1/50
Amino Acid (mg/L)	Bro	Cat	Col	Bro	Cat	Col
Alanine	0.92 (13)	0.46 (6.4)	0.82 (11.5)	0.88 (12.4)	0.29 (4.1)	1.4 (19.8)
Arginine	1.03 (6.6)	0.65 (4.2)	1.04 (6.6)	2.25 (14.4)	0.88 (5.6)	1.2 (7.7)
Asparagine	0.17 (1.5)	0.05 (0.4)	0.13 (1.2)	0.02 (0.18)	0.21 (1.8)	0.1 (0.9)
Aspartic acid	0.3 (2.5)	0.11 (1)	0.28 (2.4)	0.71 (6.1)	0.20 (1.7)	0.35 (3)
Glutamine	11.1 (76)	12.5 (85.5)	14.2 (97.2)	10.5 (71.7)	11 (75.3)	10.5 (72)
Glutamic acid	3.4 (23)	1.7 (11.4)	2.7 (18.2)	2.9 (19.4)	1.9 (13)	3.2 (21.8)
Glycine	3.4 (46)	1.9 (25.3)	3.3 (43.4)	3.8 (50.4)	3.7 (49.3)	4.1 (54)
Hydroxylysine	0.28 (1.7)	0.18 (1.1)	0.3 (1.9)	2.7 (16.7)	1.35 (8.3)	8.1 (50)
Isoleucine	0.30 (2.3)	0.13 (1)	0.24 (1.8)	0.37 (2.8)	0.08 (0.6)	0.35 (2.7)
Leucine	0.63 (4.8)	0.31 (2.4)	0.58 (4.4)	1.2 (9)	0.67 (5.1)	0.68 (5.2)
Lysine	1.61 (11)	0.34 (2.3)	0.65 (4.4)	3 (20.6)	2.33 (16)	2 (13.4)
Methionine	2.8 (18.6)	0.10 (0.6)	0.23 (1.5)	1 (7)	0.19 (1.3)	6.9 (46.3)
Phenylalanine	0.51 (3.1)	0.20 (1.2)	0.35 (2.1)	0.93 (5.6)	0.18 (1.1)	0.4 (2.5)
Proline	1.1 (9.1)	0.25 (2.1)	0.51 (4.4)	1.4 (12.5)	0.21 (1.8)	3 (25.5)
Serine	0.7 (6.7)	0.14 (1.3)	0.60 (5.7)	0.64 (6.1)	0.12 (1.15)	0.66 (6.3)
Threonine	0.8 (6.6)	0.37 (3.1)	0.67 (5.6)	0.70 (5.9)	0.55 (4.6)	0.7 (5.8)
Tyrosine	3 (16)	-	-	-	-	-
Valine	0.94 (8)	0.31 (2.6)	0.57 (4.8)	1.3 (10.8)	0.21 (1.8)	1.2 (10)
Total (mg/L)	33	19.7	27.2	34.3	24.1	44.8

). Highest amount of <3000 Da peptides was found by hydrolysis with cathepsin B at the E:C 1/20, followed by the E:C 1/50. If considering as the degree of hydrolysis to 3000 Da the amount of <3000 Da peptides on the total amount of peptides (<3000 and >3000 Da), at the 1/20 ratio it was 82% when using cathepsin, 80% when using bromelain and 73% when using collagenase; at the ratio 1/50, it was 82% with collagenase, 75% with bromelain and 73% with cathepsin.

Highest concentration of free amino acids, ca. 45 mg/L, was found with the use of collagenase at the 1/50 ratio; globally, by the 1/50 ratio a high concentration of free amino acid was observed for the three enzymes used (Table 3). The concentration of each amino acid also varied according to the enzyme and to the E:C ratio; nonetheless, whatever the enzyme and the ratio, major free amino acids found were glutamine, glycine and glutamic acid. With the ratio 1/50 significant higher amounts of lysine and hydroxylysine were found by the three enzymes used; when using collagenase a significant higher amount of proline and methionine was also found.

Either enzymes or E:C ratio did not have a noticeable effect on calcium and magnesium content although some values were statistically different. Concentration of Ca²⁺ was in the order 3860–4066 mg/L, and 220–255 mg/L for Mg²⁺ (Figure 2). All the calcium passed in the <3000 Da fraction through filtration; for magnesium the highest amount passed in the <3000 Da however it was also found in the >3000 Da (50–160 mg/L), and in particular in the collagenase and cathepsin hydrolysates. Depending on the enzyme, concentration of calcium was 21–29-fold higher than peptides; concentration of peptides was in the same order of magnesium concentration. Regarding phosphorous, by the FTIR analysis it was possible to determine that it was present (Figure 3). As expected, tri-calcium phosphate, Ca₃(PO₄)₂ (picks at the frequencies 1156, 1076, 1017 and 990 cm⁻¹, Figure 3) was predominant accounting for 75% of the total phosphates’ integration area; di-calcium phosphate, CaHPO₄ (pick at 937 cm⁻¹, Figure 3), was also found and accounting for 25% of the total phosphate’s integration area. Hence, the concentration of phosphorous, stoichiometrically calculated from the concentration of calcium, was estimated to be 2300–2350 mg/L for all the <3000 Da fractions.

3.1. Effect of the Enzyme, of the Ratio E/C and of the Treatment on the Antioxydant Activity

3.1.1. ABTS Test

Peptides-minerals complexes were able to inhibit ABTS•⁺ at different percentages according to the enzyme, the ratio and the treatment, and the antioxidant capacity increased with the reaction time (Figure 4). Collagenase B allowed for the greatest inhibition (43.4% at 180 min); the <3000 Da fraction obtained by cathepsin B showed the lowest inhibition capacity, and this was still more evident with a 3-fold concentration. For all the complexes, higher values were obtained at the E:C ratio 1/20 with respect to 1/50. A higher inhibition was found for the bromelain-collagenase B complexes mixing with respect to non-mixed fractions, albeit the increase was less than 10% (Figure 5). Steam sterilisation did not significantly affect the antioxidant capacity; 3-fold concentration of the complexes allowed doubling the antioxidant capacity whatever the enzyme used (Figure 4d). Inhibition of ABTS•⁺ significantly (Pearson’s coefficient of determination >0.85, absolute value) correlates negatively with the concentration of the peptides and positively with the concentration of free amino acids glutamic acid, serine and threonine.

3.1.2. FRAP Test

Complexes were able to reduce iron from Fe(III) to Fe(II). The reducing power increased with the time of reaction (Figure 6). The fraction obtained by cathepsin were the most effective; no significant differences were noticed between complexes obtained by bromelain and collagenase. The E:C ratio did not have a significant effect (Figure 6a,b). The reducing power significantly increased with sterilisation for the complexes obtained by cathepsin and bromelain; on the contrary, a reduction was noticed for the fraction obtained by collagenase (Figure 6c). A 3-fold concentration allowed doubling the reducing power; for the cathepsin fraction, a maximum value was observed after 1 h of reaction afterward it decreased towards stable values until 180 min. Complexes mixing had a negative effect on the reducing power, which was lower than values reported in Figure 6a for all the combinations (values not showed). The reducing power of iron correlates positively with any of the variables and negatively (<−0.8) with glutamic acid, glycine and threonine, and with the inhibition of the ABTS•⁺.

3.1.3. ORAC

Complexes obtained by collagenase had the highest ORAC value, either at the E:C 1/20 or 1/50, followed by cathepsin and bromelain complexes (Figure 7a). Mixing had a significant effect on the increase of ORAC values, which were more than triplicated when mixing collagenase and cathepsin complexes, followed by bromelain and cathepsin complexes (Figure 7b). It was any significant positive and negative correlation of ORAC values with variables; a slight positive correlation (0.66) was found with glycine. ORAC values expressed in the official measure unit are reported in

Table 4. ORAC values at the E:C ratio 1/20 and 1/50, for 100 g of solution and for 100 g of solid matter of each fraction.

	ORAC Values (Mean Values) (μmol TE/100 g)
	E:C 1/20
	Bro	Cat	Col	Bro-Cat	Bro-Col	Col-Cat	Bro-Cat-Col
100 g solution	132	147	220	639	241	731	530
100 g solid matter	20,222	19,411	19,848	95,890	96,980	95,011	95,950
	E:C 1/50
	Bro	Cat	Col	Bro-Cat	Bro-Col	Col-Cat	Bro-Cat-Col
100 g solution	126	120	188	556	171	620	470
100 g solid matter	19,940	19,582	19,800	83,226	25,740	92,483	70,401

.

3.1.4. DPPH Test

No significant antioxidant activity was noticed toward the inhibition of the radical DPPH• (Inhibition <2%, values not showed).

3.1.5. TBArs Inhibition in Meat and Butter

Peptides-minerals complexes were able to inhibit the generation of TBArs (Figure 8). Regarding meat, for sample volumes of 500, 250 and 100 μL the TBArs inhibition was of 100% for all samples, which could implicate a possible saturation phenomenon. When using lower volumes, no saturation effect was present and the best result, ca. 88% TBArs inhibition, was obtained with collagenase, at the ratio E:C 1/50 and 50 μL of sample (Figure 8). Among all the variables considered, inhibition of TBArs significantly positively correlated (>0.8) only with alanine; a positive correlation (>0.8) was also found with the ORAC value.

Regarding butter, 100% TBArs inhibtion was observed for peptide sample volumes of 500 and 250 μL; on the contrary, no effect was obtained with lower volumes of 50 and 25 μL. With the volume of 100 μL no saturation was observed and the best result was obtained with collagenase, at the E:C ratio 1/20, with ca. 77% of inhibition (Figure 9). The effect of E:C ratio was significant, unless for cathepsin. At the 1/50 ratio differences among enzymes were less important than at the 1/20 ratio. The TBArs inhibition in butter correlates, either positively or negatively, with any of the variables considered. A significant increase of the TBArs inhibition either in meat or butter was obtained by complexes mixing and in particular by the mix of the three enzymatic fractions (

Table 5. Percentage of TBArs inhibition (% I) for peptides-minerals mixing, <3000 Da fractions, 1/20 and 1/50 E:C ratios, at ambient Temperature.

	Meat (1 g) + 50 μL Sample		Butter (1 g) + 100 μL Sample
Mix	% I (1/20)	% I (1/50)	% I (1/20)	% I (1/50)
Bro-Cat	77 ± 1.02 a, A	79 ± 0.56 b, A	92 ± 0.32 a, A	87.7 ± 0.65 b, A
Bro-Col	90 ± 0.62 a, B	89 ± 0.85 a, B	95.4 ± 0.53 a, B	92.7 ± 0.95 b, B
Cat-Col	81 ± 0.92 a, C	83 ± 0.51 b, C	91 ± 0.91 a, A	90 ± 0.33 a, C
Bro-Cat-Col	93 ± 0.87 a, D	95 ± 0.21 b, D	95.8 ± 1.02 a, C	92.5 ± 0.42 b, D

For each food matrix, lowercase letters indicate differences among ratios; uppercase letters indicate differences among complexes.

). Heating to 70 °C for 2 h, and sterilisation, did not affect the inhibition capacity.

3.2. Agglometarive Hierarchical Clustering (AHC) of Complexes

The AHC analysis identified three classes (green line, blue line and black line in Figure 10) to group the six peptides-minerals complexes based on their similarities; variance was decomposed as 33% within class and 67% between classes. Complexes obtained by bromelain at both the ratios had the most similar effect, followed by complexes obtained by cathepsin. The complex obtained by collagenase at the ratio 1/50 on the contrary, shows the highest dissimilarity with respect to the 1/20 ratio and to the other four complexes.

4. Discussion

Peptides were released in a different amount according to the enzyme and to the ratio used (Figure 1). Bromelain and cathepsin B belong to the same class of enzymes and are nonspecific proteases; in spite of this similarity, bromelain preferred cutting sites are hydrophilic amino acids while large hydrophobic amino acids are the preferred cutting sites of cathepsin B [18]. Collagenase B, on the contrary, belongs to a different class of enzymes; it is a specific protease for which the preferred cutting site is a residue of glycine. The finding of the highest amount of <3000 Da peptides by using cathepsin B, at both the E:C ratios, can be likely explained by the fact that collagen is a hydrophobic protein being principally composed of hydrophobic amino acids such as glycine, proline and alanine [17]. The finding that bromelain generated higher amount of <3000 Da peptides with respect to collagenase B could be explained by the fact the total hydrophilic amino acids in collagen are present in a larger amount with respect to glycine [17]. Moreover, while bromelain and cathepsin B have a similar molecular weight and dimension, collagenase B has a significantly higher weight and dimension (Table 1), which could generate a steric impediment to access the specific cutting site [18]. This could also probably explain the finding that, by the use of collagenase, a significant higher amount of free amino acids was released at the 1/50 E:C ratio with respect to the 1/20 ratio where the molecular crowing is most important.

With respect to the capacity to quench free radicals, although the ABTS•⁺ and the APPH• (2,2′-azobis(2-methylpropioanamidine) dihydrochloride, the radical targeted in the ORAC test) can not fully represent biological molecules, their application can give practical information on the potential of collagen peptides to transfer electrons or a hydrogen atom towards a radical molecule so as to neutralise it [22]. In particular, the ABTS test can be representative of aromatic nitrogen radicals quenching [22]. This essay allows to easily comparing related classes of samples, as in the case of our study, and can reflect changes in biological and food samples that undergo treatments/processing (heating, sterilisation, etc.) [13,28]. Moreover, as the ABTS•⁺ radical is soluble in both organic and aqueous solvents [22], the essay showed that collagen peptides-minerals complexes can quench both hydrophilic and lipophilic radicals. Another aspect to consider is that the ABTS•⁺ chemical structure is very similar to the neutral tryptophan indolyl radical (TprN•); this radical plays a role in several radical pathways and in particular in the oxidation of tyrosine, an amino acid which is the precursor of dopamine and noradrenaline, thus having important functions in the human body [29]. Hence, the results of our study could also indicate the activity of collagen peptides-minerals complexes for the neutralization of the tryptophan indolyl radical. Eventually, the ABTS•⁺ inhibition positively correlated with free amino acids serine and threonine which are both hydrogen donors (and also acceptors), and this finding agrees with the mechanism of the ABTS•⁺ quenching that is based either on the transfer of hydrogen or electrons [22].

Regarding the ORAC test, values found for 100 g of peptides-minerals complexes in the liquid form, and not mixed, compare, for instance, with 100 g of carrot and courgette, and with 100 g of the herb parsley, in the case of the bromelain and cathepsin B fractions; higher ORAC values obtained with collagenase compare with 100 g of cucumber and apple vinegar [30]. Mixing of the peptides-minerals complexes allowed to increase the ORAC value, in particular in the case of collagenase B and cathepsin B < 3000 Da fractions mixing, where 100 g of peptides-minerals complexes in the liquid form compare with 100 g of tomato and with 100 g of extra-virgin oil enriched with parsley or basil [30]. If the 100 g of dry matter are considered, significant higher ORAC values were found (Table 4); they compare with spices, herbs and brans (like cocoa, cumin, dried marjoram and dried parsley, sorghum brain), which globally have significant higher ORAC values with respect to vegetables and oils.

The reducing capacity of Fe(III) to Fe(II) of the peptides-minerals complexes was significant and compared with food of plant origin [31]. Sterilization and concentration significantly improved the antioxidant activity. The highest reducing power was found for the cathepsin B fraction, a result that could be explained by the lower amount of free amino acids in this fraction and then by the lower amount of electrons acceptors amino acids (such as threonine) and bivalent amino acids (such as glycine) that can compete with Fe(III) in receiving an electron; this also likely explains the negative correlation of FRAP values with threonine and glycine. Some researches previously carried out, and in particular by Carlsen at al. [31], highlighted that plant-based food show higher FRAP antioxidant power with respect to food of animal origin or mixed plant/animal food. In this regards, the positive results of our study can then support the finding that animal proteins, such as collagen, can be a source of peptides that have higher antioxidant activity than the protein itself, and these peptides can compare with plant origin food.

When assessing the antioxidant activity towards lipids peroxidation for food such as bovine meat and butter, the peptides-minerals complexes showed interesting results that could promote their application in industry. As expected, the volume of peptides-minerals complexes to add to have a comparable TBArs inhibition in meat and butter was lower (however not linearly proportional) in the case of meat (Figure 8 and Figure 9) since the amount of lipids is lower with respect to butter. The values of the TBArs essay for meat significantly correlated with the ORAC values, which agrees with the target of both essays which are peroxyl radicals. A positive correlation was also find with butter albeit less significant. Lipid peroxidation refers to the oxidative degradation of lipids that contain in their structure carbon-carbon double bonds, as, for exemple, unsaturated fatty acids (including ω3 and ω6), phospholipids, glycolipids, etc., [32]. Malondialdehyde (MDA) is the main product of unsaturated lipid oxidation and it can negatively affect the nutritional quality of food due to the formation of adducts with food proteins [23,33]. From a biological point of view, in the human body, MDA levels increase with age and this can give rise to pro-inflammatory and pro-fibrogenic effects as a consequence of the formation of MDA-proteins and MDA-DNA (deoxyribonucleic acid) adducts. In this regard, patients affected by the Parkinson’s disease, have showed higher levels of MDA with respect to healthy individuals [33]. These finding suggest that the effect of bovine bone collagen peptides-minerals complexes could be also tested in aging and in neurodegenerative disease status such as the Parkinson’s sickness. Apart the antioxidant capacity, the peptides-minerals complexes could be also regarded as a nutritional source of calcium, phosphorous and magnesium. They are essential minerals for the growth and health of bones and teeth, and particular important in the prevention of osteoporosis; also, they are indispensable for a number of biochemical reactions such as the biosynthesis of phospholipids, DNA, adenosine thiphosphate, stabilisation of cells membranes, etc. [19,34]. Apart the importance in diet, calcium phosphates are approved as food additives (with the number E341, in Europe) to regulate the acidity of foods, to improve the texture and processability of flour, and as white pigments [34].

5. Conclusions

The three enzymes tested showed to be effective in the generation of antioxidant peptides-minerals complexes from bovine bone. Not always an improved result was noticed by an increased amount of the enzyme (E:C ratio effect); in some cases, the 1/50 ratio allowed for best results, which is positive from an economic point of view since a lower quantity of enzyme can be used. The antioxidant power depended by the enzyme and by the essay; however, the mix of the different peptides-minerals fractions can allow for a significant increase of the activity, whatever the essay. Nonspecific collagen enzymes such as bromelain and cathepsin B were effective in the generation of antioxidant peptides; however, peptides obtained by collagenase B showed the highest antioxidant activity in all the essay. Depending on the target radical, the antioxidant activity of peptides obtained by collagenase B can be still improved by mixing with peptides obtained by bromelain or cathepsin B. The results of this study confirmed that animal by-products of food-grade such as bovine bone obtained from slaughterhouses could be valorised for high-value food and health application. As a future direction, a study of the antioxidant activity of the sole mineral phase of the complexes will be carried out to understand its antioxidant power, as well as a study of the synergies between peptides and minerals toward their antioxidant action.