**Antioxidant Potential of Mung Bean (***Vigna radiata***) Albumin Peptides Produced by Enzymatic Hydrolysis Analyzed by Biochemical and In Silico Methods**

#### **Jennifer Kusumah, Luis M. Real Hernandez and Elvira Gonzalez de Mejia \***

Department of Food Science and Human Nutrition, University of Illinois at Urbana-Champaign,

Urbana, IL 61801, USA; kusumah2@illinois.edu (J.K.); rhealmlus@gmail.com (L.M.R.H.)

**\*** Correspondence: edemejia@illinois.edu; Tel.: +1-217-244-3196; Fax: +1-217-265-0925

Received: 13 July 2020; Accepted: 2 September 2020; Published: 4 September 2020

**Abstract:** The objective of this study was to investigate the biochemical antioxidant potential of peptides derived from enzymatically hydrolyzed mung bean (*Vigna radiata)* albumins using an 2,2- -Azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS) radical scavenging assay, a ferrous ion chelating assay and an oxygen radical absorbance capacity (ORAC) assay. Peeled raw mung bean was ground into flour and mixed with buffer (pH 8.3, 1:20 *w*/*v* ratio) before being stirred, then filtered using 3 kDa and 30 kDa molecular weight cut-off (MWCO) centrifugal filters to obtain albumin fraction. The albumin fraction then underwent enzymatic hydrolysis using either gastrointestinal enzymes (pepsin and pancreatin) or thermolysin. Peptides in the hydrolysates were sequenced. The peptides showed low ABTS radical-scavenging activity (90–100 μg ascorbic acid equivalent/mL) but high ferrous ion chelating activity (1400–1500 μg EDTA equivalent/mL) and ORAC values (>120 μM Trolox equivalent). The ferrous ion chelating activity was enzyme- and hydrolysis time-dependent. For thermolysin hydrolysis, there was a drastic increase in ferrous ion chelating activity from t = 0 (886.9 μg EDTA equivalent/mL) to t = 5 min (1559.1 μg EDTA equivalent/mL) before plateauing. For pepsin–pancreatin hydrolysis, there was a drastic decrease from t = 0 (878.3 μg EDTA equivalent/mL) to t = 15 (138.0 μg EDTA equivalent/mL) after pepsin was added, but this increased from t = 0 (131.1 μg EDTA equivalent/mL) to t = 15 (1439.2 μg EDTA equivalent/mL) after pancreatin was added. There was no significant change in ABTS radical scavenging activity or ORAC values throughout different hydrolysis times for either the thermolysin or pepsin–pancreatin hydrolysis. Overall, mung bean hydrolysates produced peptides with high potential antioxidant capacity, being particularly effective ferrous ion chelators. Other antioxidant assays that use cellular lines should be performed to measure antioxidant capacity before animal and human studies.

**Keywords:** albumin; albumin peptide; antioxidant peptide; bioactive peptide; in silico; mung bean; mung bean albumin; peptide sequencing; *Vigna radiata*

#### **1. Introduction**

Mung bean, also known as green gram, is a small, green-colored legume widely cultivated throughout Asia [1]. It is a popular legume in countries such as Indonesia and China where its consumption is associated with positive health outcomes [2,3]. Mung bean flour is commonly made into a paste and incorporated into bread and desserts [4]. Mung beans have a relatively high protein content (19.5–33.1%) that is comparable to that of soybeans (*Glycine max*) (35–50%) and kidney beans (*Phaseolus vulgaris*) (23–25%) [5–7]. Due to its nutritional content, mung bean can be a plant-based protein source in developing nations where animal protein sources are cost-prohibited [8]. Compared to other legumes, mung beans are relatively free from antinutritional factors [9]. Mung beans are also rich in vitamins and minerals such as iron, magnesium, potassium, copper and folate [10].

The major storage proteins of mung bean are globulins (62.0%), albumins (16.3%), glutelins (13.3%) and prolamins (0.9%), with vicilin-type protein (8S) making up 89% of globulins [11]. Globulins are the main storage proteins in mature mung beans, and they are also the most well studied mung bean proteins [12]. In contrast to globulins, there are limited studies on the albumin proteins of mung beans [13]. Albumins are water-soluble, globular proteins found in both animals and plants [14]. In plants, albumins can be proteins stored in seeds to be used during germination and growth [15]. The isolation and characterization of some albumins in multiple legumes, such as lentils, soybeans and winged beans, has already been performed [16–18]. There are currently two main entries for mung bean albumin in the UniProt Database; Q9FRT8 is a reviewed entry detailing a 10 kDa protein fragment. Q43680 is a non-reviewed entry detailing a 30 kDa protein. The sequences of these two entries are given in Figure 1, and the sequences have minimal similarity (7% identity). There is currently no evidence that only a single mung bean albumin exists, so reported mung bean albumin sequences can be different from each other. Given the current data, mung bean albumins are not expected to have a molecular weight greater than 30 kDa [19,20].


**Figure 1.** Alignment of the mung bean albumin sequences currently present in the UniProt Database. The alignment was conducted using Clustal Omega. Q9FRT8 is the reviewed sequence of a 10 kDa protein fragment, while Q43680 is the non-reviewed sequence of a 30 kDa protein. The sequences have 7% identity, with 19 identical positions and 21 similar positions. Asterisks (\*) indicate identical amino acids in the sequences, while dots indicate amino acids in the sequences that are not identical.

Peptides derived from chickpea albumins, which are legume albumins like those present in mung beans, have shown high biologically relevant antioxidant potential [21,22]. Antioxidant peptides can benefit human health by chelating excess transition metal ions and scavenging free radicals and reactive oxygen species [23]. Lunasin, a peptide derived from soybean 2S albumin, has also been found to have antioxidant effects [24,25]. Peptides from whole mung bean protein hydrolysates have been found to have calcium and ferrous ion binding activity that can have biological implications, but this activity is also useful in preventing oxidation in food systems [26]. However, the antioxidant potential of mung bean albumin hydrolysates and peptides alone, without the presence of other mung bean proteins, has been poorly studied [27]. The objective for this study was to investigate the biochemical antioxidant potential of peptides derived from enzymatically hydrolyzed mung bean albumins using either thermolysin or gastrointestinal enzymes pepsin and pancreatin, followed by sequencing and characterizing the peptides. Antioxidant potential was investigated using an ABTS radical scavenging assay, a ferrous ion chelating assay and ORAC assay.

#### **2. Materials and Methods**

#### *2.1. Materials*

Peeled and split raw mung bean dry seeds were purchased locally (Asian Best, Thailand), and stored at 4 ◦C. The purchased mung beans were from a single brand cultivated in Thailand. The beans were taken from 4 ◦C bulk storage for each experimental sampling, at least three times for each experiment performed.

Centrifugal ultrafiltration filters with 3 or 30 kDa MWCO membranes were purchased from Millipore-Sigma (St. Louis, MO, USA). Protein reagents A and B, 2× Laemmli sample buffer, <sup>10</sup><sup>×</sup> tris/glycine and 10<sup>×</sup> tris/glycine/SDS buffers, mini-PROTEAN® TGX™ gels, Coomassie blue and Precision Plus Protein™ Dual Xtra standard were purchased from Bio-Rad (Hercules, CA, USA).

Simply Blue Safe Stain was purchased from Invitrogen (Carlsbad, CA, USA). All other reagents were purchased from Sigma-Aldrich (St. Louis, MO, USA) unless stated otherwise.

#### *2.2. Mung Bean Albumin Extraction*

Peeled dry mung beans were milled and sieved using a number 35 mesh (500 μm). Collected flour was stored at 4 ◦C until use. Since a standard extraction protocol for mung bean albumins has not been established, two different legume albumin extraction methods were implemented to extract mung bean albumins.

For one method, a recently proposed protocol for the extraction of mung bean albumins was followed as specified in Du et al. [28] with specific modifications. The collected mung bean flour was mixed with double-distilled water at a ratio of 1:20 (*w*/*v*). The pH of the mixture was adjusted to either 3.0 or 7.5 with 1 M HCl or NaOH, respectively. The extraction pH values used were previously found to solubilize a higher concentration of mung bean albumins compared to total mung bean proteins [28]. Albumins were extracted for 1 h at 25 ◦C with manual stirring of each mixture every 15 min. The mixtures were then centrifuged at 11,000× *g* for 20 min at 4 ◦C, and the collected supernatants had their pH adjusted to 4.6 using 0.1 M HCl or NaOH. The supernatants at pH 4.6 were centrifuged at 11,000× *g* for 20 min at 4 ◦C, and the precipitated mung bean albumin pellets were collected. The mung bean albumin pellets for each extraction pH were solubilized individually in 5 mL of double-distilled water (pH = 11) and sonicated for 1 min to facilitate solvation. The resulting solutions were filtered using 30 kDa MWCO ultrafiltration centrifugal filters by centrifuging the solutions at 6600× *g* for 1 h at 4 ◦C. The resulting permeate was collected and used as an aqueous filtrate of extracted mung bean albumins. The collected permeates were stored at 4 ◦C for ≤3 days.

The other extraction method followed the protocol published by Singh, Rao and Singh [29] with minor adaptations. Collected mung bean flour was mixed with 0.1 M borate buffer (pH = 8.3) at a ratio of 1:20 (*w*/*v*) and was continuously stirred for 1 h. The mixture was then centrifuged at 11,000× *g* for 20 min at 4 ◦C, and the supernatant was collected. As an alternative to extensive dialysis, the collected supernatant was filtered using a 3 kDa MWCO ultrafiltration centrifugal filter at 6600× *g* for 1 h at 4 ◦C, and then the retentate was supplied with sodium citrate buffer (pH = 4.6) to replenish the buffer volume lost as permeate. The solution at pH 4.6 was centrifuged at 11,000× *g* for 20 min at 4 ◦C, and the collected supernatant was filtered using a 30 kDa MWCO ultrafiltration centrifugal filter at 6600× *g* for 1 h at 4 ◦C. The resulting permeate was collected and used as an aqueous filtrate of extracted mung bean albumins. The collected permeates were stored at 4 ◦C for ≤3 days

#### *2.3. Gel Electrophoresis*

2.3.1. Sodium Dodecyl Sulfate–Polyacrylamide Gel Electrophoresis (SDS-PAGE)

The DC protein assay was used to determine the protein concentrations of mung bean albumin samples using a bovine serum albumin standard curve. SDS-PAGE was performed as published before with minor modifications [30]. Briefly, mung bean samples were mixed with 2× Laemmli buffer (1:1 *v*/*v*) without 2-mercaptoethanol and boiled for 5 min. The prepared samples were loaded on 4–20% Tris-glycine gels so that each well had 20 μg of crude mung bean albumins (unfiltered extract) or 15 μg of filtered mung bean albumins (≤30 kDa permeate). A Tris-glycine buffer consisting of 25 mM Tris and 192 mM glycine at pH 8.3 was used in both the anode and cathode, and the gels ran at 200 V for 30 min. SimplyBlue Safe Stain was used to stain the gels.

#### 2.3.2. Native Blue–Polyacrylamide Gel Electrophoresis (Native Blue-PAGE)

The SDS-PAGE methodology mentioned above was modified to run native blue gels. Mung bean samples were mixed (1:1 *v*/*v*) with Tris-glycine buffer containing 25% glycerol. Prepared samples were loaded on 4–20% Tris-glycine gels so that each well had 20 μg of crude mung bean albumins (unfiltered extract) or 15 μg of filtered mung bean albumins (≤30 kDa permeate). Tris-glycine buffer was used in the anode, and Tris-glycine buffer containing 2% Coomassie blue dye was used in the cathode. The gels ran at 150 V for 1 h at 22 ◦C. The gels were de-stained with a 25% isopropanol solution (*v*/*v*) containing 10% acetic acid (*v*/*v*).

#### *2.4. In Silico Hydrolysis of Mung Bean Albumin Sequences*

To determine proteases, outside of those present during gastrointestinal digestion, capable of potentially producing antioxidant peptides from mung bean albumins, a manual theoretical hydrolysis of the two reported mung bean albumin sequences (Q9FRT8 and Q43680) present in the UniProt Database (https://www.uniprot.org/) was performed. The theoretical hydrolysis was carried out using the protease specificity data present in the MEROPS Database (https://www.ebi.ac.uk/merops/). Alcalase, stem bromelain, ficin, papain and thermolysin were selected for the theoretical hydrolysis as they are food-safe enzymes capable of being used in the commercial hydrolysis of mung bean albumins.

For each mung bean albumin sequence, fragments of ≤8 adjacent amino acids, corresponding to an amino acid sequence that could occupy sites P4-P4 in the active site of a protease, were matched to the possible amino acids known to be present at the P4-P4 sites of a protease when a protein substrate is hydrolyzed by that protease. The specificity of each protease analyzed is detailed in Table 1. If a fragment sequence matched the possible amino acids that would lead the sequence to be hydrolyzed by a specific protease, the corresponding P1 and P1 amino acids in the fragment sequence were color-coded, as this is where the sequence would be expected to be hydrolyzed. Amino acids in the mung bean albumin sequences were color-coded either red or purple where hydrolysis would be expected to happen, and hydrolysis was expected to be possible between adjacent amino acids in the sequence that were color-coded the same color. Possible mung bean albumin peptide sequences that would be expected to be produced by each protease were obtained by cutting each mung bean sequence between two adjacent amino acids that were color-coded the same to produce fragments of various amino acid quantities. For conciseness and applicability, only possible di- and tri-peptides that could be produced for each protease were accounted for, as these small peptides have intestinal transporters that make them more bioavailable than their larger counterparts [31]. The bioactive fragments present in the two mung bean albumin sequences hydrolyzed were identified using the Database of Bioactive Peptides—BIOPEP Database (http://www.uwm.edu.pl/biochemia/index.php/pl/biopep).


**1.** Possible Amino Acids at Positions in or Around Cleavage Site for Specific Proteases.

**Table** 

#### *2.5. Enzymatic Hydrolysis*

The simulated gastrointestinal digestion of mung bean albumins was carried out using conditions used previously by Mojica and de Mejia [32]. Briefly, pepsin was added to a solution (pH = 2) of mung bean albumins (≤30 kDa fraction) at a ratio of 1:20 *w*/*w* for 2 h at 37 ◦C. After 2 h, pancreatin was added to the solution at a pepsin–mung bean albumin ratio of 1:20 *w*/*w*. The pH of the solution was increased to 7.5 using 1 M NaOH, and the solution was incubated for another 2 h at 37 ◦C. Enzymatic reactions were stopped by heating the solution to 90 ◦C for 15 min. The inactivated hydrolysate was stored at 4 ◦C for ≤2 days.

Thermolysin hydrolysis was carried out by adding thermolysin to a mung bean albumin solution (≤30 kDa fraction) at a thermolysin–mung bean albumin ratio of 1:20 *w*/*w*. The pH of the solution was adjusted to 8.0 using 1 M NaOH, and then the mixture was incubated at 70 ◦C for 4 h. Afterwards, the solution was heated to 95 ◦C for 15 min to stop proteolytic activity. The inactivated hydrolysate was stored at 4 ◦C for ≤2 days.

#### *2.6. Antioxidant Assays*

All antioxidant assays were carried out on fresh hydrolysates produced within ≤2 days of storage before the assays were conducted. Radical scavenging and ferrous ion chelating activities were calculated based on equations obtained from the standard curves using ascorbic acid and EDTA solutions, respectively.

#### 2.6.1. ABTS Radical Scavenging Assay

The ABTS radical scavenging assay was performed as published before for a mung bean meal hydrolysate [26]. A 7 mM ABTS and 2.5 mM potassium persulphate stock solution in 10 mM PBS buffer (pH = 7.4) was made and stored in darkness for 16 h. The stock solution was then diluted with 10 mM PBS buffer until an absorbance of 0.70 ± 0.05 at 734 nm was obtained. A sample of mung bean albumins (20 μL, 1 g/L) was added to 1980 μL of diluted ABTS solution. The mung bean albumin concentration was determined by the DC protein assay using a bovine serum albumin (BSA) standard curve, and preliminary studies were performed to determine the concentration used. The reaction was allowed to react for 5 min in the dark, and the resulting absorbance was read at 734 nm. The results were expressed as ascorbic acid equivalent in μM.

#### 2.6.2. Ferrous Ion Chelation Assay

The chelating activity of ferrous ions by mung bean peptides was analyzed as published before for a mung bean meal hydrolysate [26]. Mung bean albumins (100 μL, 1 g/L) were mixed with 100 μL of a 2 mM iron chloride solution, and the mixture was diluted with 1400 μL of double distilled water. Mung bean albumin concentration was determined by the DC protein assay using a bovine serum albumin (BSA) standard curve, and preliminary studies were performed to determine the concentration used. The mixture was incubated for 3 min at 25 ◦C. Afterwards, 400 μL of a 5 mM ferrozine solution was added, and the solution was incubated for 10 min at 25 ◦C. EDTA was used as a standard. The resulting absorbance was read at 562 nm. The result was expressed as EDTA equivalent in μM.

#### 2.6.3. Oxygen Radical Absorbance Capacity (ORAC) Assay

An ORAC assay on mung bean albumin hydrolysates was performed with some modification as published before [33]. Briefly, 20 μL of the sample (1 g/L) was mixed with 120 μL of fluorescein (0.12 mM). The absorbance was then read at 485 nm and the mixture was incubated at 37 ◦C for 15 min. A total of 60 μL of 40 mM 2- 2-Azobis(2-amidinopropane) dihydrochloride (AAPH) was then added,

and the solution was read again at 582 nm. Trolox was used as standard. The results were calculated using Equations (1) and (2) and reported as Trolox equivalent (μM):

$$ALIC = \left(\frac{R\_1}{R\_1}\right) + \left(\frac{R\_2}{R\_1}\right) + \left(\frac{R\_3}{R\_1}\right) + \dots + \left(\frac{R\_n}{R\_1}\right) \tag{1}$$

$$\text{Net AlIC} = \text{AlIC}\_{\text{sample}} - \text{AlIC}\_{\text{blank}} \tag{2}$$

#### *2.7. Peptide Sequencing*

Peptide sequencing was performed as published before [34]. Briefly, peptides obtained from the different hydrolysates were analyzed by HPLC–ESI–MS/MS using a Q-ToF Ultima mass spectrometer (Waters, Milford, MA, USA) equipped with an Alliance 2795 HPLC system. The gradient mobile phase A was 95% water, 5% acetonitrile and 0.01% formic acid, while mobile phase B was 95% acetonitrile, 5% water and 0.1% formic acid. The volume of injection was 400 μL/min and the PDA detector wavelength was 280 nm. Each peak was analyzed in MassLynx V4.1 software (Waters Corp., Milford, MA, USA) and the sequence of amino acids was identified based on the accurate mass measurements, while tandem MS fragmentation using the MassBank database was used to analyze the data and obtain the peptide sequences with >80% of certainty. The isoelectric point, net charge and hydrophobicity of the peptides were analyzed by PepDraw [35]. The amino acids were presented as one letter nomenclature.

#### *2.8. Statistical Analysis*

The experiments were repeated at least three independent times from different beans, starting from the mung bean albumin extraction and proceeding up to the antioxidant assays and peptide sequencing. Data are expressed as mean ± standard deviation. The data obtained were analyzed using one-way ANOVA to compare experimental to control values, and differences were considered significant at *p* < 0.05. GraphPad Prism 7 (GraphPad Software, LLC., San Diego, CA, USA) was used for the data analysis.

#### **3. Results**

#### *3.1. Mung Bean Albumin Profiles*

The mung bean albumin extraction methodologies used in this study resulted in mung bean albumin extracts with different protein profiles (Figures 2 and 3). Only the albumin isolate obtained from mung bean albumins with borate buffer pH 8.3 had no 40–50 kDa globulins (Figure 2, lane 6). Therefore, the albumin isolate obtained from the borate buffer pH 8.3 extraction methodology was determined to be purer than the other protein isolates analyzed. Since mung bean albumins are not currently known to be larger than 30 kDa, the 30 kDa permeate of the albumin isolate obtained from the borate buffer pH 8.3 extraction protocol was sequenced. In addition, it was used for all antioxidant assays performed, as this isolate was determined to be the most consistent with the current literature [19,20]. The protein profile of this permeate is given on lane 9 of Figure 2.

**Figure 2.** SDS-PAGE (Sodium Dodecyl Sulfate–Polyacrylamide Gel Electrophoresis) of mung bean albumin extracts and 30 kDa MWCO (molecular weight cut-off) filtrates. MWL = molecular weight ladder. (**1**) Water pH = 7.5 extract, (**2**) Albumin isolate from water pH = 7.5 extract, (**3**) Water pH = 3.0 extract, (**4**) Albumin isolate from water pH = 3.0 extract, (**5**) Borate buffer pH = 8.3 extract, (**6**) Albumin isolate from borate buffer pH = 8.3 extract. Samples on lanes (**2**), (**4**) and (**6**) were filtered using a 30 kDa MWCO centrifugal filter to produce the permeates on (**7**), (**8**) and (**9**), respectively. Arrows indicate the protein fractions that are most likely to be mung bean albumin according to their molecular weight.

**Figure 3.** The native Blue-PAGE of mung bean albumin extracts and 30 kDa MWCO filtrates. BSA = bovine serum albumin. BSA was used as a reference protein. (**1**) Water pH = 7.5 extract, (**2**) Albumin isolate from water pH = 7.5 extract, (**3**) Water pH = 3.0 extract, (**4**) Albumin isolate from water pH = 3.0 extract, (**5**) Borate buffer pH = 8.3 extract, (**6**) Albumin isolate from borate buffer pH = 8.3 extract. Samples on lanes (**2**), (**4**) and (**6**) were filtered using a 30 kDa MWCO centrifugal filter to produce the permeates on (**7**), (**8**) and (**9**), respectively. Arrows indicate the protein fractions that are most likely to be mung bean albumin according to their molecular weight.

#### *3.2. Potentially Antioxidant Peptides from Mung Bean Albumins by In Silico Hydrolysis*

For the 10 kDa mung bean albumin sequence Q9FRT8, the bioactive fragments LK and FC were found for the sequence using the BIOPEP Database. These fragments were found in the Q9FRT8 sequence three times, and their locations are shown in Figure 4. For the 30 kDa mung bean albumin sequence Q43680, the bioactive fragments HL, AY, EL, IR, LK, KP, VY and FC were found using the BIOPEP Database. These fragments were found in the Q43680 sequence 11 times, and their locations are shown in Figure 5.


**Figure 4.** Locations of possible hydrolysis, by proteases selected, within the mung bean albumin sequence Q9FRT8 of the UniProt Database. Amino acids in the sequence are coded using their one letter abbreviation. Hydrolysis is expected to occur between any two adjacent amino acids color-coded the same color (red or purple), but not between amino acids color-coded with different colors or in black text. Each block represents a part of the sequence depicting the different specificities of the proteases stated in the left column. The top block is the start of the sequence. Antioxidant fragments according to the BIOPEP Database in the sequence are highlighted in blue.


**Figure 5.** Locations of possible hydrolysis, by proteases selected, within the mung bean albumin sequence Q43680 of the UniProt Database. Amino acids in the sequence are coded using their one letter abbreviation. Hydrolysis is expected to occur between any two adjacent amino acids color-coded the same color (red or purple), but not between amino acids color-coded with different colors or in black text. Each block represents a part of the sequence depicting the different specificities of the proteases stated in the left column. The top block is the start of the sequence. Antioxidant fragments according to the BIOPEP Database in the sequence are highlighted in blue.

Thermolysin was found to be the protease most likely to produce small bioactive di-/tri-peptides from both mung bean albumin sequences in the UniProt Database analyzed (Table 2). Thermolysin was also found to be the enzyme most likely to destroy bioactive fragments in the mung bean albumin sequences analyzed, but given that it could hydrolyze the majority of the mung bean albumin sequences, this was expected (Figures 4 and 5). Thermolysin was selected to hydrolyze mung bean albumins in vitro due to its greater potential to produce antioxidant peptides.


**Table 2.** Possible Small Antioxidant Peptides Produced from the Theoretical Hydrolysis of Mung Bean Albumin Sequences by Papain and Thermolysin.

<sup>1</sup> Potentially antioxidant fragments in the analyzed di-/tri-peptides are labeled in blue.

#### *3.3. Antioxidant Activity of Mung Bean Albumin Hydrolysates and Peptide Sequencing*

Results in Table 3 and Figure S1 show that mung bean albumins had some antioxidant capacity themselves (t = 0). Thermolysin and pepsin–pancreatin hydrolysis increased the antioxidant capacity in terms of ferrous ion chelating activity with significant changes due to time of hydrolysis. The ferrous ion chelating activity of the hydrolysates derived via both thermolysin and pepsin–pancreatin enzymatic hydrolysis was generally higher than the ABTS radical scavenging activity. ABTS and ORAC values did not present a significant increase in the antioxidant activities of any of the hydrolysates produced by the enzymes tested. The ferrous ion chelating activities of thermolysin and gastrointestinal peptides hydrolysates were not statistically different at the end of the hydrolysis. The same pattern was observed for ORAC values between thermolysin and gastrointestinal peptides. There was, however, significant difference between t = 0 and the rest of the time points for the ferrous ion chelating activity of thermolysin-derived peptides, except for pepsin, where the activity decreased. There was no significant difference in the ABTS radical scavenging activity at different time points for both thermolysin-derived and simulated gastrointestinal digestion-derived peptides.


**Table 3.** Antioxidant Potential of Mung Bean Hydrolysates with Varying Hydrolysis Times.

MBA = Mung Bean Albumin; 1.0 mg/mL of EDTA is equivalent to 3421.8 μM; 1.5 mg/mL of EDTA is equivalent to 5132.8 μM; 0.15 mg/mL of ascorbic acid is equivalent to 851.7 μM; 1.0 mg/mL of ascorbic acid is equivalent to 5677.9 μM.

Results in Tables 4 and 5 also show that the peptides derived from enzymatic hydrolysis using gastrointestinal enzymes (pepsin-pancreatin), and with antioxidant potential, have higher molecular mass (~328 Da) compared to peptides derived from enzymatic hydrolysis using thermolysin (~253 Da). Enzymatic hydrolysis using thermolysin also produced more peptides with high hydrophobicity (>10 kcal/mol). All the peptides reported in Tables 4 and 5 have been found to have antioxidant potential according to the BIOPEP Database, which is compounded from previously published literature.


**Table 4.** Sequences and Functional Properties of Mung Bean Albumin Peptides Derived from Simulated Gastrointestinal Digestion.



Supplementary Materials Table S1 presents the complete sequences and functional properties of mung bean albumin hydrolysates derived from gastrointestinal enzymatic hydrolysis, and the effect of thermolysin. The values of the peptides found ranged between 236 and 1509 kDa, 5.81 and 20.74 Kcal/mol and 2.82 and 11.18 for molecular mass, hydrophobicity and isoelectric point, respectively. Supplementary Materials Table S2 presents the peptide sequences, functional properties and bioactivities of mung bean albumin hydrolysates derived from thermolysin enzymatic hydrolysis.

#### **4. Discussion**

A defining characteristic of mung bean albumins is their high solubility in water, but extracting them with water alone was found to be inadequate to isolate them from other mung bean proteins, as demonstrated by the SDS-PAGE result in Figure 2. The extraction of mung bean albumins using sodium borate buffer (pH = 8.3) and then subsequent centrifugal filtration was more effective for obtaining isolated mung bean albumins that were within the currently known filtration 3–30 kDa mass range. At pH 7.0, other mung bean proteins aside from albumins were soluble [28], and therefore the extraction of mung bean albumin with water alone at pH 7.0 will inevitably induce the extraction of other proteins. The solubility of other mung bean proteins lowered as the pH became higher than 7.0, while the solubility of albumin increased as the pH neared 8.0. Thus, using a buffer of pH 8.3 to extract mung bean albumins was more effective than using pure water, as the difference in solubility between albumin and the rest of the proteins at that pH allowed mung bean albumins to be effectively isolated. Mung bean albumin peptides produced through hydrolysis by thermolysin and gastrointestinal enzymes showed antioxidant potential. Mung bean albumin peptides showed high ferrous ion chelating activity, but low ABTS radical scavenging activity (Supplementary Materials Figure S1). The ferrous ion chelating activity of mung bean albumin peptides derived from

pepsin–pancreatin enzymatic hydrolysis was found to be higher than the ferrous ion chelating activity of cowpea and common bean protein hydrolysates derived from the same enzyme scheme, as reported by Segura-Campos et al. [36]. The ferrous ion chelating activity of mung bean albumin peptides was also found to be higher than the ferrous ion chelating activity of soybean lunasin, as reported by Garcia-Nebot et al. [37]. It was also found to have higher ferrous ion chelating activity compared to the Bambara groundnut (*Vigna subterranea*) protein hydrolysates, as reported by Arise et al. [38]. It also has a much higher ferrous ion chelating activity compared to phaseolin and bean protein hydrolysates, as reported by Carrasco-Castilla et al. [39] However, it has lower ferrous ion chelating activity compared to pea (*Pisum sativum* L.) protein hydrolysates, as reported by Pownall et al. [40] Thus, it can be said that mung bean albumin peptides are effective ferrous ion chelators. Of note, food-safe commercial enzymes such as alcalase and flavourzyme have been reported to produce hydrolysates with higher ferrous ion chelating activity compared to pepsin–pancreatin, as demonstrated by the sources above. These results are similar in this study, wherein thermolysin was found to produce peptides with higher iron ion chelating activity compared to pepsin–pancreatin, although the difference was also found to be not statistically significant.

Ascorbic acid was used as a standard in the ABTS assay as it is a water-soluble antioxidant present in foods. Previous publications also report the use of ascorbic acid as a standard for the ABTS assay [41].

The iron chelating and radical scavenging activities showed variation at different hydrolysis times for peptides derived via thermolysin and those derived via pepsin–pancreatin enzymatic hydrolysis (Figure S1). Overall, mung bean albumin peptides produced through hydrolysis by thermolysin showed higher radical scavenging and metal chelating activity, although the difference was not statistically significant at the end of hydrolysis. However, both types of peptides showed similar ORAC values.

According to the study on mung bean protein hydrolysates done by Sonklin et al. [42], radical scavenging activity was affected by both enzyme concentration and the hydrolysis time of a hydrolysate. Increasing enzyme concentration and hydrolysis time increased radical scavenging activity, but only until reaching a critical point, after which the activity became constant and decreased. As we used the same concentration of enzyme throughout the hydrolysis process, the variation in antioxidant capacity can be explained by the difference in hydrolysis time. Budseekoad et al. [26] reported that the iron ion binding capabilities of mung bean protein hydrolysates vary with hydrolysis time, and that different enzymes produce peptides with the highest calcium ion binding capabilities at different times. It can be theorized that different peptides with different antioxidant capacities were produced at different hydrolysis times, as was observed by this investigation.

Studies performed by Kong and Xiong [43] and Ajibola et al. [44] revealed that hydrophobicity played a part in the antioxidant capacity of peptides, with more hydrophobic peptides exhibiting a higher antioxidant capacity. Enzymatic hydrolysis using thermolysin produced mostly peptides with high hydrophobicity (>10 kcal/mol) compared to enzymatic hydrolysis using gastrointestinal enzymes, which explains the difference in the antioxidant capacity between the two types of hydrolysates analyzed. According to Zhu et al. [45], the size of peptides is a significant factor in their antioxidant capacity, whereby peptides with lower molecular mass exhibit higher antioxidant capacities. This assessment agreed with our results, as the thermolysin enzymatic hydrolysis produced peptides with lower molecular mass (Table 5).

Both high hydrophobicity and small molecular weight are factors that increase the absorption of peptides in the small intestine [46]. Peptides produced with thermolysin were of small molecular weight and of high hydrophobicity, indicating that they are expected to be absorbed during digestion. Antioxidant and hydrophobic casein peptides have already been shown to have good bioavailability using a Caco-2 cell model [47]. Future cellular studies are needed to measure the antioxidant capacity of mung bean protein-derived peptides before animal and human studies are performed.

#### **5. Conclusions**

Mung bean albumins at pH 7.0 have the same solubility as other mung bean proteins, making their extraction using pure water ineffective in isolating mung bean albumins from other proteins. Mung bean albumins at pH 8.3 had higher solubility than other mung bean proteins, making this pH better for mung bean albumin extraction. Mung bean albumin hydrolysates showed antioxidant potential in terms of ferrous ion chelating and ORAC values. They are particularly effective ferrous ion chelators. Hydrolysates produced via thermolysin enzymatic hydrolysis had higher antioxidant capacity overall due to their high hydrophobicity and low molecular mass. High hydrophobicity and low molecular mass are two factors that can increase the intestinal absorption of thermolysin-derived mung bean albumin peptides, but more in vivo studies are required to quantify their bioavailability in humans. The variation of antioxidant capacity over different time points during hydrolysis showed that different hydrolysis times produced different peptides of different antioxidant capacities. To our knowledge, this is the first study to investigate the antioxidant potential of mung bean albumin peptides through a variety of methods.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2304-8158/9/9/1241/s1, Table S1: Sequences, Functional Properties and Bioactivity of Mung Bean Albumin Hydrolysates Derived from Gastrointestinal Enzymatic Hydrolysis; Table S2: Sequences, Functional Properties and Bioactivity of Mung Bean Albumin Hydrolysates Derived from Thermolysin Enzymatic Hydrolysis; Figure S1: Antioxidant Potential of Mung Bean Albumin Hydrolysates at Various Hydrolysis Times from Two Different Enzyme Schemes. The star marks the time point at which pancreatin was added 2 h after pepsin. These results were obtained from hydrolysates derived from an average of 576 μg/mL of mung bean albumin.

**Author Contributions:** Conceptualization, J.K., E.G.d.M.; Methodology, L.M.R.H., J.K.; Validation L.M.R.H., J.K.; Formal Analysis, J.K., L.M.R.H.; Investigation, J.K.; Resources, E.G.d.M.; Data Curation, J.K., L.M.R.H.; Writing—original draft preparation, J.K., L.M.R.H.; Writing—review and editing, J.K., L.M.R.H., E.G.d.M.; Visualization, J.K., L.M.R.H., E.G.d.M.; Supervision, E.G.d.M.; Project Administration, E.G.d.M.; Funding acquisition, J.K., E.G.d.M. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by USDA-NIFA 1014457 to EDM and the Undergraduate Research program of College of Agricultural, Consumer and Environmental Sciences to JK, University of Illinois at Urbana-Champaign.

**Acknowledgments:** The authors thank Diego Luna-Vital for his help during the course of this study and Erick Damian Castañeda Reyes for his help in the data analysis section.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **In Silico and In Vitro Analysis of Multifunctionality of Animal Food-Derived Peptides**

#### **Lourdes Amigo 1, Daniel Martínez-Maqueda <sup>2</sup> and Blanca Hernández-Ledesma 1,\***


Received: 6 July 2020; Accepted: 21 July 2020; Published: 24 July 2020

**Abstract:** Currently, the associations between oxidative stress, inflammation, hypertension, and metabolic disturbances and non-communicable diseases are very well known. Since these risk factors show a preventable character, the searching of food peptides acting against them has become a promising strategy for the design and development of new multifunctional foods or nutraceuticals. In the present study, an integrated approach combining an in silico study and in vitro assays was used to confirm the multifunctionality of milk and meat protein-derived peptides that were similar to or shared amino acids with previously described opioid peptides. By the in silico analysis, 15 of the 27 assayed peptides were found to exert two or more activities, with Angiotensin-converting enzyme (ACE) inhibitory, antioxidant, and opioid being the most commonly found. The in vitro study confirmed ACE-inhibitory and antioxidant activities in 15 and 26 of the 27 synthetic peptides, respectively. Four fragments, RYLGYLE, YLGYLE, YFYPEL, and YPWT, also demonstrated the ability to protect Caco-2 and macrophages RAW264.7 cells from the oxidative damage caused by chemicals. The multifunctionality of these peptides makes them promising agents against oxidative stress-associated diseases.

**Keywords:** bioactive peptides; animal protein; multifunctionality; antioxidant activity; in silico; cell models

#### **1. Introduction**

Non-communicable diseases (NCDs) such as cardiovascular and neurodegenerative disorders, cancer, and diabetes, are the principal cause of death and disability worldwide [1]. Most of these diseases are caused by environmental factors, with the diet being one of the main contributing factors. While the consumption of highly processed foods and sugar-sweetened beverages has been associated with a higher risk of these disorders, a healthy diet including functional foods may help in reducing or even preventing several NCDs [1,2]. Thus, in recent years, the search for bioactive food compounds and their use as substitutes of pharmacological treatments has intensified. Due to their desirable impacts on human health and limited side effects, bioactive peptides have become one of the most studied food components, being usually included into functional foods and nutraceuticals [3]. Once liberated from the source protein by enzymatic hydrolysis, gastrointestinal digestion, or food processing, bioactive peptides may act on different body systems exerting different functionalities such as antihypertensive, antioxidant, opioid, antithrombotic, hypocholesterolemic, anticancer, immunomodulatory, and antimicrobial activities, among others [4]. Moreover, it has been demonstrated that some food peptides are able to exert two or more bioactivities, acting on several systems at the same time [5]. This has made it so that multifunctional peptides have been

recently recognized as more useful than peptides with single activity as they influence multiple cell processes, affecting different signaling pathways simultaneously [6]. Among food sources of multifunctional peptides, milk and meat proteins are considered some of the most studied [7]. As examples, caseinophosphopeptides have been reported to exert anticariogenic, antihypertensive, immune-enhancing, antigenotoxic, and cytomodulatory effects [3]. Whey protein lactoferrin and its derived peptide, lactoferricin, are well known by their anticancer, antitumor, immunomodulatory, and antimicrobial activities [8].

The rising evidence suggests a possible common pathophysiology among NCDs, with oxidative stress and hypertension as the main contributing factors [9,10]. Oxidative stress occurs when reactive oxygen species (ROS) overload the body's defenses or when these defenses lose their capacity to react, leading to damage of essential cell components [11]. Experimental, clinical, and epidemiological studies have revealed that this status is involved in the development of NCDs such as arteriosclerosis, obesity, type 2 diabetes, inflammatory bowel disease, arthritis, neurological, liver, and renal disorders, and cancer [12]. Hypertension, defined as high blood pressure, is currently considered one of the major preventable risk factors linked to cardiovascular diseases [13].

The classical or empirical approach, also referred to as the in vitro method, is the most employed approach in peptide bioactivity screening. However, it requires exhaustive sample preparation and does not always allow the explicit identification of particular bioactive peptides. To overcome these major disadvantages, bioinformatics-driven (in silico) approaches have recently been developed. These strategies enable estimating potential precursor proteins through the calculation of quantitative descriptors, constructing profiles of the potential biological activity of peptide sequences, and predicting peptidic bonds susceptible to enzymatic hydrolysis [14]. Thus, in silico analyses have been recognized as playing a significant role in the process of bioactive peptides generation and identification [15]. However, there are many limitations associated with bioinformatic data, such as the extent, quality, and reliability of published information within databases. Moreover, the validation of attributed bioactive sequences, as well as the study of other aspects such as peptides' stability, bioavailability, and mechanisms of action should be carried out to complete the provided information by in silico analyses. Therefore, different milk and meat-derived peptides that are similar to or share amino acids with previously described opioid fragments were selected and subjected to an integrated approach combining an in silico study and in vitro assays to evaluate their multifunctionality and mechanisms of action as basis of their future use in functional foods.

#### **2. Materials and Methods**

#### *2.1. Materials*

Triisopropyl silane, angiotensin-converting enzyme (ACE), fluorescein (FL), 2,2- -azinobis(3 ethylbenzothiazoline-6-sulfonic acid) diammonium salt (ABTS), 3-(4,5-dimethylthiazol-2-yl)-2,5 diphenyl tetrazolium bromide (MTT), dimethylsulfoxide (DMSO), dichlorofluorescin (DCFH), and Hank's Balanced Salt Solution (HBSS) were purchased from Sigma-Aldrich (St. Louis, MO, USA). 2,20-azobis (2-methylpropionamide) dihydrochloride (AAPH) and 6-hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid (Trolox) were obtained from Aldrich (Milwaukee, WI, USA). Abz-Gly-Phe (NO2)-Pro was purchased from Bachem Feinchemikalien (Bubendorf, Switzerland), and trifluoroacetic acid (TFA) was obtained from Scharlau (Barcelona, Spain). High-Glucose Dulbecco's Modified Eagle Medium (DMEM), fetal bovine serum (FBS), and penicillin/streptomycin/amphotericin B solution were purchased from Biowest (Kansas City, MO, USA). A 1% non-essential amino acids solution was from Lonza Group Ltd. (Basilea, Switzerland).

The milk and meat-derived peptides used in this study (Table 1) were synthesized by the conventional Fmoc solid-phase synthesis method using an Applied Biosystems model 433A synthesizer (Foster City, CA, USA). The cleavage of the peptides from the polystyrene-based resin (Applied Biosystems, Foster City, CA, USA) was carried out with TFA, triisopropyl silane, and MilliQ water for 2 h. The analysis of peptides was carried out by high performance liquid chromatography coupled to tandem mass spectrometry (HPLC-MS/MS) analysis using an Agilent 1100 HPLC System (Agilent Technologies, Waldbron, Germany) connected online to an Esquire 3000 ion trap (Bruker Daltonik GmbH, Bremen, Germany) and equipped with an electrospray ionization source. The identification of peptides and their purity were determined through mass comparison and peak integration, respectively (Supplementary Figure S1). Finally, peptides were dissolved in 10% acetic acid, freeze-dried, and kept at −20 ◦C until further analysis.

#### *2.2. Peptide Screening by In Silico Analysis*

Synthetic peptides were subjected to in silico analysis by using the Milk Bioactive Peptide Database (MBPDB [16] and BIOPEP-UWM database of bioactive peptides (BIOPEP-UPW) [17]).


**Table 1.** Source protein, fragment, sequence, molecular mass, and purity of synthetic peptides used in the present study.

The potential of peptides to be bioactive was predicted using PeptideRanker software, and their theoretical bioactivity was expressed as score values calculated (from 0 to 1, with 1 being the most likely to be bioactive). Moreover, prediction of the toxicity was performed using ToxinPred.

#### *2.3. Angiotensin Converting Enzyme (ACE)-Inhibitory Activity*

The ACE-inhibitory activity of synthetic peptides was determined by the fluorescence protocol optimized by Sentandreu and Toldrá (2006) [18] and modified by Quirós et al. (2009) [19]. Briefly, the substrate Abz–Gly–Phe(NO2)–Pro was dissolved (0.45 mM) in 150 mM Tris and 1125 mM NaCl buffer (pH 8.3) and maintained at 4 ◦C until its use. ACE (1 U/mL) was diluted (0.04 U/mL) in 150 mM Tris buffer containing 0.1 μM ZnCl2 (pH 8.3). Forty microliters of sample (or MilliQ water for blank and control) were added to a black multi-well plate (Porvair, Leatherhead, UK). Then, 40 μL of ACE were added, and the reaction started after the addition of 160 μL of the substrate. The plate was

incubated at 37 ◦C for 30 min, and the fluorescence was measured in a FLUOstar OPTIMA plate reader (BMG Labtechnologies GmbH, Offenburg, Germany) with 320 nm excitation and 420 nm emission filters. Data were processed with the FLUOstar Control version 1.32 R2 (BMG Labtech) software and expressed as IC50 (peptide concentration needed to inhibit 50% of the ACE activity).

#### *2.4. In Vitro Antioxidant Activity*

#### 2.4.1. Oxygen Radical Absorbance Capacity (ORAC)-FL Assay

An oxygen radical absorbance capacity (ORAC)-FL assay was used based on the protocol previously optimized [20]. The reaction was performed at 37 ◦C in 75 mM phosphate buffer (pH 7.4). The final assay mixture volume was 200 μL, containing 70 nM FL, 12 mM AAPH, and antioxidant [(Trolox, 1–8 μM) or sample (at different concentrations)]. Fluorescence was measured during 137 min in a FLUOstar OPTIMA plate reader (BMG Labtech) with 485 nm excitation and 520 nm emission filters. The equipment was controlled by the FLUOstar Control ver. 1.32 R2 software for fluorescence measurement. Samples were analyzed in triplicate. The final ORAC-FL value was expressed as μmol Trolox equivalents (TE) per μmol peptide.

#### 2.4.2. ABTS Assay

Antioxidant activity was measured using a previously optimized method [21] with some modifications. A mixture of 7 mM ABTS stock solution and 2.45 mM potassium persulfate was kept in the dark at room temperature for 12-16 h to form the ABTS radical cation (ABTS•+). The ABTS•<sup>+</sup> solution was diluted in 5 mM of phosphate buffer solution (PBS) (pH 7.4) to an absorbance of 0.70 ± 0.02 at 734 nm at 30 ◦C. Two mL of diluted ABTS•<sup>+</sup> solution were mixed with 20 μL of sample or Trolox (0–0.015 μmol), and the absorbance was recorded at 734 nm after 10 min incubation at 30 ◦C. The Trolox equivalent antioxidant capacity (TEAC) value was expressed as μmol TE per μmol peptide. Each sample was analyzed in triplicate.

#### *2.5. Cell Culture*

The human colorectal adenocarcinoma Caco-2 and the mouse macrophage RAW 264.7 cell lines were obtained from the American Type Culture Collection (ATCC, Rockville, MD, USA). Caco-2 and RAW264.7 cells were grown in High-Glucose DMEM supplemented with 10% (v:v) FBS, 1% (v:v) penicillin/streptomycin/amphotericin B, and 1% (v:v) non-essential amino acids solution. Cells were maintained in plastic 75-cm2 culture flasks at 37 ◦C in a humidified incubator containing 5% CO2 and 95% air.

#### *2.6. Cell Treatment Conditions*

Cells were incubated for 24 h with various concentrations of selected synthetic peptides. To evaluate both the direct and protective effects against oxidative stress, the incubation period was followed by 1.5 h-treatment with culture medium or *tert*-butyl hydroperoxide (*t*-BOOH, 1 mM for Caco-2 cells and 0.25 mM for RAW264.7 cells), respectively, and different biomarkers were evaluated.

#### 2.6.1. Cell Viability

Cell viability was determined using the MTT assay. Caco-2 and RAW264.7 cells were seeded onto 96-well plates (Corning Costar Corp., Corning, NY, USA) at a density of 1.0 <sup>×</sup> 104 cells/cm2 and incubated for 9 days and 24 h, respectively. Then, cells were washed with PBS, treated with synthetic selected peptides (1–100 μM), and incubated for 24 h. Afterwards, culture medium was removed, and the cells were washed with PBS and incubated with medium (direct effects) or chemical oxidant (protective effects) at 37 ◦C for 1.5 h. At the end of the treatment time, 100 μL of MTT solution (0.5 mg/mL final concentration) were added to each well, and the plate was incubated for 2 h at 37 ◦C. The supernatant was aspirated, the formazan crystals were solubilized in DMSO:ethanol (1:1, v:v),

#### *Foods* **2020**, *9*, 991

and the absorbance was measured at 570 nm in a FLUOstar OPTIMA plate reader (BMG Labtech). Results were expressed as percentage of the control, considered as 100%. Samples were analyzed in triplicate.

#### 2.6.2. Determination of Intracellular Reactive Oxygen Species (ROS)

The intracellular ROS levels were detected using the ROS-sensitive fluorescent dye, DCFH, as previously described [22]. Caco-2 cells were plated in 48-well plates (density of 4.75 <sup>×</sup> <sup>10</sup><sup>4</sup> cells/well) and RAW264.7 cells in 24-well plates (density of 2.0 <sup>×</sup> 105 cells/well) and incubated for 7 days and 14 h, respectively. After this time, cells were treated with peptides as previously described for 24 h. Then, DCFH was dissolved in HBSS and added to cells at a final concentration of 5 μM solution. Cells were incubated for 30 min at 37 ◦C. The probe was removed, and cells were incubated with PBS (direct effects) or *t*-BOOH (protective effects) for 60 min, measuring the fluorescence after 60 min in a FLUOstar OPTIMA plate reader (BMG Labtech) with 485 nm excitation and 520 nm emission filters. The results were expressed as percentage of the control, which was considered as 100%. The assay was run in triplicate.

#### *2.7. Statistical Analyses*

All data were analyzed from three independent experiments. Results were expressed as the mean ± standard deviation (SD). Data were statistically analyzed by performing a one-way ANOVA test, followed by Tukey's multiple comparison test with the IBM SPSS Statistics for Windows 23.0 (IBM Corporation, Armonk, NY, USA). A *p*-value of less than 0.05 was considered statistically significant. Significant differences of each concentration versus the control under the same experimental conditions were expressed by \*\*\* (*p* < 0.001), \*\* (*p* < 0.01), and \* (*p* < 0.05).

#### **3. Results and Discussion**

#### *3.1. In Silico Analysis of Synthetic Peptides*

The physicochemical characteristics and predicted toxicity and biological activity of synthetic peptides are shown in Table 2. The bioactivity of peptides was predicted using the PeptideRanker program. There were differences in the theoretical bioactivity of peptides, with score values from 0.2453 to 0.9558. Eleven peptides, released from β-casein (β-CN), β-lactoglobulin (β-Lg), α-lactalbumin (α-La), and β-hemoglobin (β-Hg), were to be highly bioactive with a predicted bioactive score over 0.80. None of the 27 peptides were considered toxic, according to the software ToxinPred. The results of the in silico analysis by using the MBPDB and BIOPEP-UWM databases are shown in Table 3. Of the 27 synthetic peptides studied, 23 were already included into databases because of their opioid, ACE-inhibitory, antioxidant, anticancer, antidiabetic, or immunostimulating activities, among others. Some of them were found to exert two or more activities. Thus, αs1-CN-derived peptides RY, RYL, RYLG, YLGY and FYPEL exert both ACE-inhibitory and antioxidant activities [23–27]. ACE-inhibitory and opioid activities are exerted by sequences YGFLP [28] and YGFL [29,30]. Four of the analyzed peptides (RYLGY, LGY, YPFPGPI, and YLLF) have been reported to exert three or more activities. Among them, the multifunctional β-casein A2-derived peptide YPFPGPI is highlighted by exerting ACE and dipeptidyl peptidase IV (DPP-IV) inhibitory, anticancer, anxiolytic, immunomodulatory, opioid, antidiabetic, and satiating activities (Table 3). Only four peptides (YPFPGPIP, YPFVEP, YGFL, and YPW) were found to be novel, although they share active sequences with bioactive peptides released from the same source protein.


**Table 2.** Physicochemical characteristics and predicted biological activity and toxicity of synthetic peptides derived from food sources (PeptideRanker and ToxinPred databases).

<sup>1</sup> pI: Isoelectric point.

**Table 3.** Predicted biological activity of synthetic peptides derived from food sources using Milk Bioactive Peptide Database (MBPDB) and BIOPEP-UWM database of bioactive peptides.



**Table 3.** *Cont.*

\* According to BIOPEP-UWM database; \*\* According to Milk Bioactive Peptide Database (MBPDB); IC50 a: Values (μM) are given for peptide concentrations inhibiting the angiotensin-converting enzyme (ACE) activity by 50%; IC50 b: Values (μM) is given for peptide concentration inhibiting (3H)-dihydromorphine binding, instead of (3H)-naloxone, by 50%; IC50 c: Values (μM) are given for peptide concentrations inhibiting (3H)-naloxone binding by 50%; Stimulated lymphocyte proliferation d: % stimulation (+) and inhibition (–), respectively, compared to control; IC50 e: Values (μM) are given for peptide concentration required to inhibit 50% of dipeptidyl peptidase IV (DPP-IV); n.d. No available data. ORAC: oxygen radical absorbance capacity.

#### *3.2. In Vitro ACE Inhibitory and Antioxidant Activities of Synthetic Peptides*

The measured ACE-inhibitory and antioxidant activities of assayed peptides are shown in Table 4. Our study confirmed the ACE-inhibitory activity already reported for several of the analyzed fragments. Moreover, this effect was newly found in other sequences such as YLGYLE, YFYPE, YPFPGPIPN, YPFVEP, YGFL, and YLL. Potent activity (IC50 values lower than 10 μM) was determined for αs1-CN peptide YFYPEL (IC50 = 8.82 μM) and β-CN peptide YPFVEP (IC50 = 7.48 μM). These values were similar to those reported for well-known milk derived tripeptides VPP (IC50 = 9 μM) and IPP (IC50 = 5 μM) [59]. The presence of leucine, valine, and proline at the C-terminus could contribute to the high ACE-inhibitory activity shown by these two sequences, as ACE prefers substrates/competitive

inhibitors with hydrophobic amino acids and/or proline at the three C-terminal positions [60]. Similarly, milk peptide HLPLP, containing leucine and proline at the C-terminus, showed an IC50 value of 41 μM [28]. As shown in Table 4, all peptides but sequence LLF presented antioxidant activity through a dual mechanism of action, hydrogen atom transfer (ORAC), and non-competitive electron transfer (ABTS). Our results confirmed the peroxyl radical scavenging activity already reported for eight of these peptides. Moreover, this property was newly found in 18 peptides whose ORAC values ranged from 1.09 ± 0.02 μmol TE/μmol peptide to 3.50 ± 0.02 μmol TE/μmol peptide. The highest value was determined for peptide YPW. The presence of tyrosine and tryptophan could determine its potent activity, as these two amino acids have been reported to be the main contributors to the peroxyl radical scavenging activity of food-derived peptides [20]. Moreover, the situation of these residues at the peptide chain could also influence the activity. Thus, when threonine was added to the C-terminus of peptide YPW, the ORAC value was reduced up to 3.19 ± 0.18 μmol TE/μmol peptide. This influence was also observed for peptides YLG and LGY. For peptide LG, the antioxidant behavior of the resultant peptides from the addition of tyrosine as a terminal residue was different. Thus, if this amino acid (Y) was added to the C-terminus, the ORAC value of peptide LGY was 2.00 ± 0.09 μmol TE/μmol peptide, while it was 0.93 ± 0.08 μmol TE/μmol peptide when tyrosine was added to the N-terminus (peptide YLG). When tyrosine was added at both termini, the ORAC value increased up to 2.96 ± 0.20 μmol TE/μmol peptide (Table 4). Our results confirmed the previously described importance of peptides parameters such as the their amino acid composition, sequence, and length in determining their antioxidative potential [61].


**Table 4.** ACE-inhibitory activity (expressed as μM) and antioxidant activity (expressed as μmol Trolox equivalents (TE)/μmol peptide of synthetic animal-protein derived peptides.

<sup>1</sup> ACE: Angiotensin-converting enzyme; <sup>2</sup> TE: Trolox equivalents; TEAC: Trolox equivalent antioxidant capacity; n.d. Activity not determined. \* ACE-inhibitory activity not detected at the highest peptide concentration analyzed (1000 μM). **\*\*** Antioxidant activity not detected at the highest peptide concentration analyzed (0.20 μmol).

Unlike peroxyl radical scavening activity, none of the analyzed peptides had been previously reported to exert ABTS radical scavenging properties. TEAC values ranged from 0.73 ± 0.07 μmol

TE/μmol peptide to 5.96 ± 0.35 μmol TE/μmol peptide (Table 4). As it has been described for the ORAC assay, the presence of tyrosine and tryptophane was responsible for the ABTS radical scavenging properties of analyzed peptides [21]. Among peptides whose antioxidant activity has been described for the first time in the present study, six sequences showed potent effects with ORAC and TEAC values higher than 2.0 μmol TE/μmol peptide. These sequences corresponded to αs1-CN peptides RYLGYLE, YLGYLE, and YFYPEL, β-CN peptide YGFLP, and β-Hg peptides YPW and YPWT. Five of these peptides were known by their opioid activity [31,32,36,37,42]. Additionally, anticancer and ACE-inhibitory activities have been reported for peptides RYLGYLE [33] and YGFLP [28], respectively. Thus, the antioxidant activity described in the present study would increase the functionality of these food-derived peptides. As a result of their multifunctionality allowing peptides to exert beneficial effects on different body systems, sequences RYLGYLE, YLGYLE, YFYPEL, and YPWT were selected to study in depth the mechanism of action involved in their antioxidant activity.

#### *3.3. Antioxidant Activity of Synthetic Peptides in Cell Models*

Two cell models, human colon adenocarcinoma Caco-2 and murine macrophages RAW264.7, were used to evaluate the protective effects of animal protein-derived peptides on the cell oxidative status under normal and chemical-induced conditions. The action of peptides on cell viability and ROS generation was studied. The direct effects of peptides RYLGYLE, YLGYLE, YFYPEL, and YPWT on Caco-2 and RAW264.7 cells viability were evaluated using the MTT assay. This assay provides a sensitive measurement of the metabolic status of the mitochondria, which reflects early cellular redox changes [62]. Treatment of Caco-2 and RAW264.7 cells with synthetic peptides did not evoke changes in cell viability, indicating that the concentrations selected (1–100 μM) did not damage cell integrity during the 24-h period of incubation.

To study the protective effects of peptides against chemical-induced oxidative damage in Caco-2 cells and macrophages, they were pre-incubated with peptides for 24 h, exposed to *t*-BOOH for 1.5 h, and then, cell viability was measured. As shown in Figure 1A–D, treatment of Caco-2 cells with *t*-BOOH (1 mM) provoked a significant reduction of cell viability of 25%, compared to non-stimulated cells.

In a previous study in our lab, García-Nebot et al. (2014) [22] had reported a 20% reduction of the Caco-2 cells' viability after 1.5-h treatment with 3 mM *t*-BOOH. A longer incubation time (6 h) with this chemical at concentrations of 0.1 and 4 mM has been also found to cause significant reductions of cell viability [63,64]. Similarly, the viability of macrophages was significantly reduced after treatment with *t*-BOOH (0.25 mM) by up to 44% (Figure 2A–D). Recent studies have also reported significant decreases (39%) of cell viability after treatment of RAW264.7 cells with 1 mM *t*-BOOH for 3 h [65]. Pre-treatment of Caco-2 cells with tested peptides before induction with *t*-BOOH for 1.5 h did not exert any protection from the effects of this chemical. In the case of macrophages, only treatment with peptide YPWT resulted in an increase of cell viability at concentrations between 25 and 100 μM. Thus, the percentage of viable cells increased from 70.97% (stimulated cells) to 79.58% (stimulated cells pre-treated with 25 μM YPWT) (Figure 2D). These results are in agreement with previous studies carried out with soybean protein-derived peptides that only found significant protection on *t*-BOOH-induced RAW264.7 cells at high doses [66].

**Figure 1.** Dose-dependent effects of synthetic animal-protein derived peptides (**A**) RYLGYLE, (**B**) YLGYLE, (**C**) YFYPEL, and (**D**) YPWT on cell viability of stressed Caco-2 cells with 1 mM tert-butyl hydroperoxide (*t*-BOOH). Cells were pre-treated with peptides at concentrations ranged from 1 to 100 μM for 24 h. Results were expressed as the percentage of viable cells compared to control, which was considered as 100% (% control, mean ± standard deviation (SD), n = 3). Different letters indicate significant differences (*p* < 0.05; Tukey multiple comparison test).

**Figure 2.** Dose-dependent effects of synthetic animal protein-derived peptides (**A**) RYLGYLE, (**B**) YLGYLE, (**C**) YFYPEL, and (**D**) YPWT on cell viability of stressed macrophages RAW264.7 with 0.25 mM tert-butyl hydroperoxide (*t*-BOOH). Cells were pre-treated with peptides at concentrations that ranged from 10 to 100 μM for 24 h. Results were expressed as the percentage of viable cells compared to control, considered as 100% (% control, mean ± standard deviation (SD), n = 3). Different letters indicate significant differences (*p* < 0.05) and \*\* (*p* < 0.01); \* (*p* < 0.05) significant differences of each concentration versus control under the same experimental conditions (one-way ANOVA followed by Tukey's multiple comparison test).

In order to understand the potential mechanism of cytoprotective action exerted by peptides, the intracellular ROS generation was evaluated in normal cells and cells exposed to *t*-BOOH after pre-treatment with synthetic peptides for 24 h. The direct evaluation of intracellular ROS is recognized

as a good indicator of the oxidative damage to living cells [67]. In our study, measurement of the intracellular ROS levels was carried out using DCFH as a fluorescent probe that once added to intact cells, crosses cell membranes and is oxidized to highly fluorescent dichlorofluorescein (DCF) in the presence of ROS [68]. In Caco-2 cells under normal conditions, four peptides caused a significant reduction of ROS levels at all concentrations used (Figure 3A,C,E,G).

**Figure 3.** Dose-dependent effects of synthetic animal protein-derived peptides (**A**,**B**) RYLGYLE, (**C**,**D**) YLGYLE, (**E**,**F**) YFYPEL, and (**G**,**H**) YPWT on reactive oxygen species (ROS) production in non-stressed Caco-2 cells (**A**,**C**,**E**,**F**) and Caco-2 cells stressed with tert-butyl hydroperoxide (*t*-BOOH, 1 mM). Cells were pre-treated with peptides at concentrations that ranged from 1 to 100 μM for 24 h. Results were expressed as the percentage of ROS levels compared to control, which were considered as 100% (% control, mean ± standard deviation (SD), n = 3. Different letters indicate significant differences (*p* < 0.05) and \*\*\* (*p* < 0.001); \*\* (*p* < 0.01); \* (*p* < 0.05) indicate significant differences of each concentration versus control under the same experimental conditions (one-way ANOVA followed by Tukey's multiple comparison test).

In the case of fragments RYLGYLE and YLGYLE, the highest reduction (19.3% and 17.3%) was observed after treating cells with low concentrations of these peptides (1 and 10 μM, respectively), compared to non-treated cells. However, these two peptides at 10 μM provoked an increase of ROS levels in non-stressed RAW264.7 (Figure 4A,C) cells. These results indicated that in addition to the peptide dose, the type of cell line could influence the antioxidant activity of the peptides.

The chemical exposition of Caco-2 cells to t-BOOH for 1.5 h significantly increases ROS levels (untreated Caco-2 cells 100.00 ± 4.15%; treated Caco-2 cells with 1 mM t-BOOH 230.19 ± 13.61%) (*p* < 0.05) (Figure 3B,D,F,H). The pre-treatment with animal protein-derived peptides at all assayed doses for 24 h significantly neutralized the ROS-generating ability of the chemical, but no dose-dependence was observed. The highest reduction (≈68.5%) was observed for cells treated with peptide YLGYLE at 1 and 10 μM, which had ROS levels of 175.06 ± 11.13% and 174.44 ± 11.60%, respectively, compared to non-peptide-treated cells (Figure 3D). In the case of macrophages, the exposition to *t*-BOOH resulted in a higher increase of ROS levels (untreated RAW264.7 cells 100.00 ± 4.77%; treated RAW264.7 cells with 0.25 mM *t*-BOOH 501.27 ± 44.20%) (Figure 4B,D,F,H). This result was similar to that found previously by Indiano-Romacho et al. (2019) [66]. Tested peptides did not show any antioxidant activity in macrophages. No significant ROS level decreases versus control were observed after treatment with peptides both in untreated and treated RAW264.7 cells with *t*-BOOH. This tendency is in disagreement with that observed both in the Caco-2 cell model and in the in vitro antioxidant activity measured by ORAC-FL and ABTS assays; besides that, cell viability is not compromised because of the MTT results. Among cases that can help appreciate these differences, fragment YFYPEL shows one of the most remarkable activities in the stressed Caco-2 cell model, significantly lowering ROS levels (from 231.58 ± 18.40% up to 168.88 ± 17.80% at 1 μM) and showing significant in vitro antioxidant activity (2.66 ± 0.16 and 2.59 ± 0.17 μmol TE/μmol peptide for ORAC and TEAC, respectively), but without any antioxidant activity in homologous conditions in a RAW264.7 cell model. It would be interesting to evaluate the suitability of this cell model for the study of antioxidant capacity.

**Figure 4.** *Cont*.

**Figure 4.** Dose-dependent effects of synthetic animal-protein derived peptides (**A**,**B**) RYLGYLE, (**C**,**D**) YLGYLE, (**E**,**F**) YFYPEL, and (**G**,**H**) YPWT on ROS production in non-stressed macrophages RAW264.7 (**A**,**C**,**E**,**F**) and macrophages RAW264.7 stressed with tert-butyl hydroperoxide (*t*-BOOH, 0.25 mM). Cells were pre-treated with peptides at concentrations that ranged from 10 to 100 μM for 24 h. Results were expressed as percentage of ROS levels compared to control, which was considered as 100% (% control, mean ± standard deviation (SD), n = 3. Different letters indicate significant differences (*p* < 0.05) and \*\*\* (*p* < 0.001); \*\* (*p* < 0.01); \* (*p* < 0.05) significant differences of each concentration versus control under the same experimental conditions (one-way ANOVA followed by Tukey's multiple comparison test).

#### **4. Conclusions**

In summary, our results have demonstrated the multifunctionality of different bioactive peptides by an approach combining in silico and in vitro assays. By the in silico analysis, different activities such as ACE inhibitory, antioxidant, opioid, and anticancer, among others, were found to be exerted by most of analyzed peptides.

Moreover, four novel peptides (YPFPGPIP, YPFVEP, YGFL, and YPW) had not been previously defined as bioactive peptides. The ACE-inhibitory and the antioxidant activities of peptides mediated through a dual mechanism of action were confirmed by in vitro assays. Four of these peptides, RYLGYLE, YLGYLE, YFYPEL, and YPWT, were selected by their potent activity, which was confirmed in the gut epithelial cell model Caco-2, protecting cells from the oxidative damage caused by chemical agents. Future animal models would be required to confirm the multifunctionality of food-derived peptides and their protective capacity against oxidative stress-associated diseases.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2304-8158/9/8/991/s1, Figure S1: RP-HPLC-MS/MS analysis of synthesized peptide YPFVE at concentration of 0.2 mg/mL (A) UV chromatogram obtained at wavelength of 214 nm; (B) Peak integration of UV chromatogram; (C) Total ion current chromatogram; (D) Average mass spectrum; (E) MS/MS spectrum of selected ion m/z 654.3 (retention time 38.3 min); (F) MS/MS fragment ions identification in sequence YPFVE.

**Author Contributions:** Conceptualization, L.A. and B.H.-L.; methodology, D.M.-M.; investigation, D.M.-M.; writing—original draft preparation, L.A., D.M.-M., and B.H.-L.; writing-review and editing, L.A.; D.M.-M.; funding acquisition, L.A., B.H.-L. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by project AGL2015- 66886-R from the Spanish Ministry of Science and Innovation (MICIU).

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

#### *Article*
