Biological activities and peptidomic profile of in vitro-digested cow, camel, goat and sheep milks

1 The present study was designed to compare in vitro digestibility, selected biological activities 2 (antioxidant, angiotensin-converting enzyme (ACE)-inhibitory and dipeptidyl-peptidase-IV (DPP-3 IV)-inhibitory activities) and digested products of proteins from skimmed cow, camel, goat and 4 sheep milks. The experimental approach combined the recently developed harmonized in vitro 5 INFOGEST digestion model and mass spectrometry to identify peptides. Goat milk had the highest 6 digestibility, while sheep milk showed the highest ACE-inhibitory activity after digestion. Cow 7 milk was found to have the highest DPP-IV-inhibitory activity. A total of 522 peptides were 8 identified after in vitro digestion of milks. Goat and sheep milk showed the highest similarity in 9 peptide sequence with 151 common peptides. Thirteen, forty-three and twenty peptides with 10 previously demonstrated antioxidant, ACE-inhibitory and DPP-IV-inhibitory activities were found 11 in digested milks. Nineteen bioactive peptides in common were released from the different milks. 12 Despite the limitations related to the analysis of one sample of milk for each species, possible 13 differences in physiological functions after the ingestion of milk from different species are 14 suggested by our results, however this requires confirmation by in vivo testing.


Introduction
Bioactive peptides have been defined as specific protein fragments that have a positive impact on body functions or conditions and may ultimately influence health (Rizzello et al., 2016).These peptides are inactive within the sequence of the parent protein and can be released under proteolytic conditions such as those in the gastro-intestinal tract or during food processing.These bioactive peptides potentially carry out their activity in the human body after the digestion process and once they are released from their original structure, and may act as regulatory compounds with hormonelike activity (Nongonierma & FitzGerald, 2015).The beneficial health effects of bioactive peptides include antimicrobial, antioxidative, dipeptidyl peptidase-IV (DPP-IV) and angiotensin-converting enzyme (ACE) inhibition, antihypertensive and immunomodulatory activities (Nongonierma & FitzGerald, 2015;Rizzello et al., 2016).Today, milk proteins are considered an important source of bioactive peptides and an increasing number of them have been identified in milk protein hydrolysates and fermented dairy products (Hernández-Ledesma, García-Nebot, Fernández-Tomé, Amigo, & Recio, 2014;FitzGerald, Murray, & Walsh, 2004;Nongonierma & FitzGerald, 2015;Egger, & Ménard, 2017).
Besides the well-known and most commonly consumed cow milk, a high consumption of milk of different origins (e.g.camel, goat and sheep milk) can be observed in other areas such as Asia, Africa and many eastern European countries.These alternative milks show high biological values, similar to those of cow milk, and are also used in the production of infant formulas or as a milk allergy-alternatives for those who suffer allergic reactions to cow milk (El-Agamy, Nawar, Shamsia, Awad, & Haenlein, 2009;Yadav, Singh, & Yadav, 2016).
Casein concentration is different between the different types of milk, whereas sheep milk has the highest concentration among cow, camel and goat milk (Park, Juárez, Ramos, & Haenlein, 2007).
Moreover, the incidence of the four major caseins (αS1-, αS2-, β-, and κ-caseins) is also different and related to the milk type (Tagliazucchi, Shamsia, Helal, & Conte, 2017).Divergence in the primary structure of milk proteins across species may have an impact on the potential bioactivities of the released peptides.
The main bioactive peptides studied are those with antioxidant, ACE-inhibitory and DPP-IV inhibitory activities (Nongonierma & FitzGerald, 2015;Hernández-Ledesma, García-Nebot, Fernández-Tomé, Amigo, & Recio, 2014).In most cases, the active peptides were released by hydrolysis with individual proteases, such as pepsin, trypsin, papain, thermolysin or combination, or through the action of microbial enzymes during milk fermentation (Rizzello et al., 2016;Abd El-Salam, & El-Shibiny, 2017).Some recent studies addressed the release of bioactive peptides after in vitro digestion (Rutella, Solieri, Martini, Tagliazucchi, 2016;Tagliazucchi, Shamsia, & Conte, 2016a;Egger, & Ménard, 2017;Tagliazucchi et al., 2017); however, there is a lack of information about the comparison between the bioactivities and he release of bioactive peptides from milks of different species after in vitro digestion.In addition, studies found in literature were focused on the release of ACE-inhibitory peptides and on determination of ACE-inhibitory activity of digested milks.For example, two recent studies applied the harmonized in vitro digestive system to study the release and fate of some ACE-inhibitory peptides, such as VPP, IPP, VY, HLPLPL during cow milk digestion (Kopf-Bolanz et al., 2014;Rutella et al., 2016).In two additional studies, 17 and 20 bioactive peptides with ACE-inhibitory activity were found in camel and goat milk, respectively, subjected to the harmonized in vitro digestion (Tagliazucchi et al., 2016a and2017).Moreover, a comparative analysis of the peptidomic profile of peptides released during in vitro digestion of different milk has never been reported until now.
Therefore, the present study was designed to compare in vitro digestibility, biological activities (antioxidant, ACE-inhibitory and DPP-IV-inhibitory activities) and digested products of proteins from skimmed cow, camel, goat and sheep milk employing a harmonized basic static in vitro digestive model, simulating human digestion and developed within the COST Action INFOGEST.

Materials
All MS/MS reagents were from Bio-Rad (Hercules, CA, U.S.A.).Chemicals and enzymes for the digestion procedure, ACE and DPP-IV assays, antioxidant activity measurements and degree of hydrolysis determination were purchased from Sigma-Aldrich (Milan, Italy).Amicon Ultra-4 regenerated cellulose filters with a molecular weight cut-off of 3 kDa were supplied by Millipore (Milan, Italy).The whole milk from camel, goat and sheep were obtained from farms at El-Alamin and Sidi-Barani areas around Alexandria (Egypt).Cow whole milk was obtained from a local producer (Reggio Emilia, Italy).All the other reagents were from Carlo Erba (Milan, Italy).

Chemical analysis of skimmed cow, camel, goat and sheep milks
Skimmed cow, camel, goat and sheep milks were prepared as reported in Tagliazucchi et al. (2017) and analysed for pH, fat, and lactose by phenol-sulphuric acid method, and total nitrogen, noncasein nitrogen by micro-Kjeldahl (Tagliazucchi et al. 2016a).Three analytical replicate for each milk sample were run for each assay.

In vitro gastro-intestinal digestion of skimmed cow, camel, goat and sheep milks using the harmonized protocol
For the in vitro digestion, the protocol previously developed within the COST Action INFOGEST was followed (Minekus et al., 2014) with minor modifications for adaptation to milk (Tagliazucchi, Helal, Verzelloni, Bellesia, & Conte, 2016b).The protocol consisted of three consecutive steps: oral, gastric and intestinal phases.The three steps were carried out in absence of light.Simulated salivary, gastric, and intestinal fluids (SSF, SGF and SIF) (Kopf-Bolanz et al., 2012) were employed for each step.First, oral digestion was performed by adding 12 mL of the stock SSF solution and 150 U mL -1 of porcine α-amylase to 9 mL of skimmed milk.The sample was shaken for 5 min at 37°C.Second, the gastric digestion step was carried out by adding to the bolus 24 mL of SGF.The pH was adjusted to 2.0 with 6 mol L -1 of HCl and supplemented with porcine pepsin (1115 U mL -1 of simulated gastric fluid).After 2 h of incubation at 37°C, the final intestinal step was carried out by adding 36 mL of SIF (prepared by mixing 24 mL of pancreatic fluid and 12 mL of bile salts).Then, the pH was adjusted to 7.0, supplemented with pancreatin and the samples were incubated at 37°C for 2 h.All samples were immediately cooled on ice and frozen at -80°C for further analysis.The digestions were performed in triplicate.In addition, a control digestion, which included only the gastro-intestinal juices and enzymes, and water in place of milk, was carried out to consider the possible impact of the digestive enzymes in the subsequent analysis.For each digestion, aliquots were taken after 0 and 5 minutes of salivary digestion, after 30, 60 ,90 and 120 minutes of gastric digestion and after 30, 60 ,90 and 120 minutes of intestinal digestion.

Assessment of protein hydrolysis during the digestion and preparation of the peptidic fractions from digested cow, camel, goat and sheep milks
Protein hydrolysis during the in vitro digestion was followed by measuring the amounts of released amino groups using the 2,4,6-trinitrobenzenesulfonic acid (TNBS) assay and leucine as standard (Adler-Nissen, 1979).The obtained raw data were corrected by the contribution of the control digestion and normalised with respect to the initial content in proteins of the respective milk.
Data are expressed as mmol leucine equivalent g -1 milk proteins and reported as a mean value and standard deviation from the three analytical replicates.Low molecular weight peptides were extracted by ultrafiltration (cut-off 3 kDa) from the post-pancreatic digested samples as described by Tagliazucchi et al. (2017).The peptide content in the peptidic fraction was determined by using the TNBS method as described above and expressing the results as mg of leucine equivalent mL -1 .

Antioxidant activities analysis
The antioxidant activity of the sample collected during the in vitro digestion procedure was determined using the 2,2′-azino-bis(3-ethylbenzothiazoline-6-sulphonic acid (ABTS) method as described in Re et al. (1999).The antioxidant properties of the peptidic fractions were evaluated using three different assays.
The ABTS assay was carried out as described above.The capacity to scavenge hydroxyl radicals was evaluated according to Tagliazucchi, Helal, Verzelloni, & Conte (2016c).In the assay, 50 μL of appropriately diluted samples or standard (vitamin C) were mixed with 50 μL of TPTZ (2,4,6-tri(2pyridyl)-S-triazine) at a concentration of 3 mmol L −1 , 50 μL of 3 mmol L −1 FeSO4, and 50 μL of 0.01% (v/v) hydrogen peroxide, in a clear bottom 96-well plate.The mixture was incubated for 1 h at 37°C, and the absorbance was measured at 540 nm using a microplate reader.
The ability to inhibit lipid peroxidation was carried out using a linoleic acid emulsion system (Tagliazucchi et al., 2016c).For that purpose, 200 μL of sample (at a peptide concentration of 1g L −1 ) were added to 200 μL of ethanol and 2.6 μL of linoleic acid, and the total volume was adjusted to 500 μL with sodium phosphate buffer, 50 mmol L −1 , and pH 7.0.The mixture was incubated at 40°C in the dark for a week.The amount of generated lipid hydroperoxide was measured by the FOX assay as reported by Tagliazucchi et al. (2010).
The obtained raw data were corrected by the contribution of the control digestion and normalised with respect to the initial content in proteins of the respective milk or to the peptide content in the peptidic fractions.ABTS scavenging capacity was expressed as μmol of vitamin C g -1 milk proteins or μmol vitamin C g -1 of peptides.Hydroxyl radical scavenging capacities was expressed as μmol vitamin C g -1 of peptides.The lipid peroxidation inhibitory activity of the samples was expressed as percentage of inhibition with respect to a control reaction carried out in presence of the peptidic fraction of the control digestion.
Three analytical replicate were run for each sample in all the assays.
For the calculation of the IC50 value, the ACE assay was carried out in presence of different amounts of the milk peptidic fractions and the data were corrected for the contribution of the control digestion.IC50 was defined as the concentration of peptides required to inhibit 50% of the enzymatic activity and expressed as μg of peptides mL -1 .The IC50 values were determined using nonlinear regression analysis and fitting the data with the log (inhibitor) vs. response model generated by GraphPad Prism 6.0 (GraphPad Software, San Diego, CA, USA).For the enzymatic assay three analytical replicate were carried out.

Measurements of dipeptidyl peptidase IV (DPP-IV)-inhibitory activity
The enzyme DPP-IV was extracted from rat intestinal acetone powder.Namely, 100 mg of intestinal acetone powder was added to 3 mL of 0.1 mol L -1 Tris-HCl pH 8.0 buffer and sonicated in a sonic bath (for 30 sec 4 times).After centrifugation at 10000g for 30 min, the resulting supernatant was directly analysed.For the calculation of the DPP-IV activity of the rat intestinal acetone extract, variable amounts of the extract (from 5 to 40 μL) were added to 5 μL of the substrate glycine-proline-p-nitroanilide (Gly-Pro-pNA 6.4 mmol L -1 ) and the 0.1 mol L -1 Tris-HCl pH 8.0 buffer was added to reach 300 μL (final volume of the assay).After 10 min of incubation at 37°C, the amount of release p-nitroanilide (pNA) was measured at 405 nm using a microplate reader.One unit of DPP-IV is defined as the quantity of enzyme that releases 1.0 μmol of pNA from Gly-Pro-pNA per minute at pH 8.0 at 37°C.
For the inhibition assay, in a 96-well plate 50 μL of diluted peptidic fractions, 235 μL of 0.1 mol L -1 Tris-HCl pH 8.0 buffer and 10 μL of enzyme solution (0.1 U mL -1 ) were added.The reaction was initiated by the addition of 5 μL of substrate solution (Gly-Pro-pNA 6.4 mmol L -1 ).After 20 min of incubation at 37°C, the amount of release p-nitroanilide (pNA) was measured at 405 nm using a microplate reader.
The concentration of peptides required to cause 50% inhibition of the DPP-IV activity (IC50) was determined by plotting the percentage of DPP-IV inhibition as a function of sample final concentration (natural logarithm).IC50 values were expressed as mg of peptides mL − 1 .Data were corrected for the contribution of the control digestion.For the enzymatic assay three analytical replicate were carried out.

Analysis of the peptidomic profile of peptidic fractions of cow, camel, goat and sheep milks by nanoflow liquid chromatography accurate mass quadrupole time-of-flight mass spectrometry with electrospray ionization (LC-ESI-QTOF MS)
The peptidic fractions from digested cow, camel, goat and sheep milks were subjected to QTOF MS/MS analysis for peptide identification.Nano LC/MS and tandem MS experiments were performed on a 1200 Series Liquid Chromatographic two-dimensional system coupled to a 6520 Accurate-Mass Q-TOF LC/MS via a Chip Cube Interface (Agilent Technologies, Santa Clara, CA, USA).Chromatographic separation was performed on a ProtID-Chip-43(II) including a 4 mm 40 nL enrichment column and a 43 mm × 75μm analytical column, both packed with a Zorbax 300SB 5 μm C18 phase (Agilent Technologies).
For peptide identification, a non-targeted approach already optimized for the analysis of digested milk was applied as reported by Tagliazucchi et al. (2016b).The mass spectrometer was tuned, calibrated and set with the same parameters as reported by Dei Più et al. (2014).This approach suffers of several limitations especially related to the detection and identification of small peptides and any peptide containing free cysteine (Fricker, 2015).Small peptides (<500 Da) are often inefficiently ionized giving a low intensity m/z signal which hampered the selection of precursor for successive MS/MS fragmentation.To overcome this problem, each digested milk was run twice by changing the range of precursor selection.In the first run MS/MS level experiments were acquired using a 4 amu precursor selection width and m/z 500-1700 scan range.To detect also small peptides, in the second run MS/MS level experiments were acquired using a 4 amu precursor selection width and m/z 50-500 scan range.The database search approach also has its limitations.
First, if the correct fragment is not derived from one of the proteins in the database, the search cannot provide the correct peptide identification.Secondly, the software commonly used for proteomic study and adapted for peptide identification, such as Mascot, have normally a minimum peptide length for identification of five residues and are not able to identify short peptides (Koskinen, Emery, Creasy, & Cottrell, 2011).Therefore, for the identification of peptides, we used a de novo sequencing software, which is able to identify also shorter peptide such as di-or tripeptides.

Statistical analysis
All data are presented as mean ± standard deviation (SD) for three replicates for each prepared digestion.Univariate analysis of variance (ANOVA) with Tukey post-hoc test was applied using GraphPad Prism 6.0 (GraphPad Software, San Diego, CA, USA).The differences were considered significant with P<0.05.

Comparison between the digestibility of cow, camel, goat and sheep milk proteins
The chemical composition of skimmed cow, camel, goat and sheep milks is reported in Table 1.
Sheep milk contained significant higher (P<0.05)amount of total proteins and caseins respect to the other milks.The content in total proteins and caseins was not significant different between cow, camel and goat milks.Indeed, no significant statistical differences were observed between the total whey proteins and lactose content as well as the pH value of the different milk.
The degradation of milk proteins by gastro-intestinal proteolytic enzymes was compared by measuring the amount of released free amino groups using TNBS assay (Figure 1).As expected, the amount of free amino groups before the digestion (corresponding to the time 0 of the salivary phase of digestion) was not significantly different between the different milk and remained constant during the 5 minutes of salivary incubation.An increase in the hydrolysis was observed for milk of different species during gastric digestion.After 30 minutes of gastric digestion the amount of free amino groups released from goat milk was significantly higher (P <0 .0 0 1 ) than that released from cow, camel and sheep milk.No significant statistical differences were observed between the milk from sheep and camel, whereas cow milk showed significantly less amino groups that the other milks (P>0 .0 5 ).The amount of released amino groups increased slightly but not significantly during the subsequent 90 minutes of peptic digestion in all the milks.The transition from gastric to pancreatic environment produced a significant increase in the amount of free amino groups in all the digested milks.Subsequently, the quantity of released amino groups showed a tendency to gradually increase during the entire pancreatic phase of the digestion.At the end of the digestion, goat milk showed a significant higher amount of released amino groups (P <0 .0 0 1 ) compared to camel, cow and sheep milks.No significant differences were found between the amount of free amino groups released from cow, camel and sheep milk (P>0 .0 5 ).These results showed that gastric and duodenal enzymes degraded goat milk proteins faster and more efficiently than camel, cow and sheep milk.These conclusions are supported by comparison with previously published data.For example, Almaas et al. (2006) found that goat milk proteins were degraded faster than cow milk using human gastro-intestinal proteolytic enzymes.On the other hand, Salami et al. (2008) found that the extent of hydrolysis of camel caseins with pancreatic enzymes was greater than that of cow caseins.Digestion of camel, cow and goat milk with the same protocol used in this study resulted in a higher digestibility of goat milk respect to camel and cow milk (Rutella et al., 2016;Tagliazucchi, et al., 2016a;Tagliazucchi et al., 2017).
The different enzyme-to-substrate ratio during the digestion, especially in the case of sheep milk, which showed the highest initial protein content, may have had an impact on the hydrolysis of milk proteins.Espejo-Carpio, Pérez-Gálvez, Guadix and Guadix ( 2013) reported an increase in the digestibility of goat milk proteins as a function of the enzyme-to-substrate ratio.The lower digestibility of sheep milk proteins can be partially attributed to the lower enzyme-to-substrate ratio respect to the other digested milks.

Evolution of antioxidant activity during in vitro digestion and antioxidant properties of the postpancreatic peptidic fractions
The variation in antioxidant activity during the digestion of the different milk was followed by the ABTS assay and reported in Figure 2.
All the studied milk showed ABTS radical scavenging activity before the digestion (corresponding to the time 0 of the salivary phase of digestion), but with some differences (Figure 2).Sheep milk had a significant higher ABTS radical scavenging activity with respect to the other milks (P<0.05),whereas cow milk showed the lowest ABTS radical scavenging activity.Clausen, Skibsted, & Stagsted (2009) found that caseins are quantitatively the highest radical scavengers in milk whereas the lower contribution of the low molecular weight compounds is due to ascorbate and especially urate.Caseins have a high content of antioxidative amino acids such as tyrosine, tryptophan and phosphoserine, and quenching of free radicals by oxidation of these amino acids was proposed as the explanation (Clausen et al. 2009).As expected, the ABTS radical scavenging activity of the different milk remained constant during the 5 minutes of salivary incubation, whereas the ABTS radical scavenging activity rose as the digestion proceeded reaching the highest value at the end of the pancreatic phase of the digestion in all the analysed milks (Figure 2).This can be explained by an increased number of peptides and amino acids at higher hydrolysis available for interaction with the ABTS radical as already reported (De Gobba, Espejo-Carpio, Skibsted, & Otte, 2014; Kumar, Chatli, Singh, Mehta, & Kumar, 2016).On the other hand, previous studies reported an increase in radical scavenging activity of cow, goat and human milk after in vitro digestion (Tsopmo, et al., 2009;Nehir et al., 2015;Power Grant et al., 2016;Tagliazucchi et al., 2016c).Comparison of the data at the end of the digestion showed that sheep and goat milk displayed the highest ABTS radical scavenging activity (P>0.05)followed by cow (P<0.001) and camel (P<0.001)milk.
Hernández-Ledesma, Amigo, Recio and Bartolomé (2007) found that an equimolar free amino acids mixture had low antioxidant activity compared to those of the corresponding peptides.Accordingly, extensive hydrolysis, resulting in an increased amount of free amino acids, should bring about to a lower antioxidant activity.However, digested sheep and goat milks showed the highest ABTS radical scavenging activity but goat milk showed the highest digestibility whereas sheep milk the lowest.Therefore, the ABTS radical scavenging activity of digested milk seems more related to the specificity and amount of formed peptides than to the extent of hydrolysis.
To fully characterize the antioxidant properties of the digested milk and to evaluate the impact of the released peptides, peptidic fractions were further extracted from the post-pancreatic digested samples through ultrafiltration with a cut-off of 3 kDa and evaluated for their ABTS radical scavenging activity and for their ability to scavenge hydroxyl radical and to inhibit lipid peroxidation.The data regarding the antioxidant properties of the peptidic fractions of the postpancreatic samples are reported in Table 2, together with the peptide content.The amount of released peptides after pancreatic digestion was not significantly different between cow, camel and goat milk whereas sheep milk digestion resulted in a release of significantly greater amount of peptides.Normalizing the data for the peptide content, it was possible to compare the antioxidant capacity of the peptidic fractions of the different milks.All of the peptidic fractions exhibited a certain degree of ABTS and hydroxyl scavenging activity.ABTS radical scavenging activity of the peptidic fractions of sheep, goat and cow milk was not significantly different whereas camel milk peptidic fraction showed the lowest ABTS radical scavenging activity (Table 2).Peptidic fraction from cow milk was the most active against hydroxyl radical whereas fractions from goat and sheep milk showed the highest lipid peroxidation inhibitory activity (Table 2).The distinct antioxidant properties of the gastro-intestinal digested peptidic fractions should be mainly attributed to the specificity of the peptides released from the sequences of the protein present in the different milk.

ACE-inhibitory activity of the post-pancreatic peptidic fractions
The ACE-inhibitory activity obtained for the peptidic fractions of the post-pancreatic samples were expressed as IC50 (defined as the peptide concentration required to inhibit 50% of the ACE activity) and ranged from 625.4 ± 60.6 to 2396.5 ± 135.0 μg of peptides mL -1 (Table 2).The hydrolysates produced by the action of digestive enzymes on sheep milk exhibited the highest ACE inhibitory activity whereas cow milk peptidic fraction showed the lowest inhibitory activity (Table 2).
The different enzyme-to-substrate ratio in the case of sheep milk could have partially influenced the ACE-inhibitory activity of the peptidic fraction of digested sheep milk.Enzymatic hydrolysis can generate ACE-inhibitory peptides whereas further degradation of the peptides into much smaller fragments may result in a decrease in the ACE-inhibitory activity (Tagliazucchi et al., 2017).
Therefore, the lower digestibility of sheep milk could result in a lower amount of short peptides and a highest ACE-inhibitory activity.Previous reported data showed that the digestion of camel and goat milk, using the same harmonized in vitro model and the same ACE assay, resulted in an IC50 value comparable with that found in this study (Tagliazucchi et al., 2016a;Tagliazucchi et al., 2017).

DPP-IV-inhibitory activity of the post-pancreatic peptidic fractions
Digests from cow, camel, goat and sheep milk showed DPP-IV inhibitory activity in the in vitro assay (Table 2).A dose dependent inhibition was observed for all digests but some differences were noted.Cow milk post-pancreatic peptidic fraction had the lowest IC50 value against DPP-IV (6.9 ± 0.1 mg peptides mL -1 ), which means the highest inhibitory activity.The other digested milks showed a DPP-IV inhibitory power from 2.2 to 2.5 times lower than cow milk, with digested camel milk having a significant lower inhibitory activity than digested goat milk.
The different DPPIV-inhibitory activity of the digested milks is probably related to differences in the amount and/or type of released peptides.However, at least in the case of sheep milk, the relatively low DPP-IV inhibitory potency may be partially linked with the lowest enzyme-tosubstrate ratio, which resulted in a lower extent of hydrolysis respect to the other milks (Nongonierma, Mazzocchi, Paolella, & FitzGerald, 2017a).

Peptidomic profile of in vitro digested cow, camel, goat and sheep milk peptidic fractions and identification of antioxidant, ACE-inhibitory and DPP-IV-inhibitory peptides
The nano-LC-MS/MS system identified 522 peptides from the digested samples.In particular, 119, 76, 164, and 163 peptides were identified in digested cow, camel, goat, and sheep milk, respectively (see online supplementary Tables S1-S8).This work reveals higher numbers of peptides released after in vitro digestion of camel and goat milk than previously reported using the same harmonized protocol.Our previous research identified 65 and 50 peptides in digested camel and goat milk, respectively (Tagliazucchi et al., 2016a;Tagliazucchi et al., 2017).Concerning cow milk, several studies have already found higher amount of peptides (more than 119) than this study (Egger et al., 2016;Picariello et al., 2010).To the best of our knowledge, this is the first paper reporting a comprehensive peptidomic profile of digested sheep milk.
The majority of the peptides were from caseins (71.4,73.7, 72.0 and 71.2% of the total identified peptides in digested cow, camel, goat and sheep milk, respectively) with β-casein which was the best source of peptides in all the digested milk (43.7, 51.3, 42.7 and 40.5% of the total identified peptides in digested cow, camel, goat and sheep milk, respectively).Whey proteins gave a lower amount of peptides respect to caseins, especially in camel milk that does not contain β-lactoglobulin (see online supplementary Tables S1-S8).In addition, 9 amino acids were also identified, 7 of them being essential amino acids (W, L, I, V, K, R and F).
The Venn diagram (Figure 3A) showed that 26, 35, 5, and 8 peptides were specific for in vitro digested cow, camel, goat, and sheep milk, respectively.Only 26 identified peptides were common for all the four digested milk, whereas goat and sheep milk showed the highest similarity in peptide sequences with 151 common peptides.Among them, 81 were in common also with cow milk and 33 with camel milk, whereas 63 peptides were found only in goat and sheep digested milk.
Tables 3-5 display the identified peptides with previously reported antioxidant, ACE-inhibitory and DPP-IV-inhibitory activities.In this study, 26 identified bioactive peptides are from β-casein, 8 from αS1-casein, 4 from αS2-casein and 4 from κ-casein.Only 3 bioactive peptides were released from whey proteins (two from β-lactoglobulin and one from α-lactalbumin).Finally, 19 peptides ranging from two to three amino acids arose from various milk proteins.
The Venn diagram (Figure 3B) showed that 19 identified bioactive peptides were common for all the four digested milks.The cow milk was the one that gave the highest number of unique bioactive peptides (8 specific peptides), whereas goat and sheep milk still showed the highest similarity in bioactive peptide sequences with 48 common peptides.
Three amino acids and 13 peptides with previously reported antioxidant properties were identified in the peptidic fraction of digested milk (Table 3).Some peptides such as VY and LK were found in the peptidic fractions of all the digested milk whereas others peptides were found only in specific fractions.In general, the majority of peptides with previously reported ABTS radical scavenging activity were found in digested goat and sheep milk, which showed the highest ABTS radical scavenging activity.On the contrary, camel milk peptidic fraction showed the lowest ABTS radical scavenging activity and contained the lowest number of ABTS radical scavenging peptides (Tables 2 and 3).Three free amino acids (tryptophan, tyrosine and phenylalanine) with previously reported antioxidant properties were also identified in all the peptidic fractions of digested milk.These amino acids had been previously suggested as the major contributors to the antioxidant activity of digested human and cow milk (Tsopmo et al., 2009;Tagliazucchi et al., 2016c).In general, the presence of an antioxidant amino acid seems to be fundamental for the antioxidant properties of a peptide (Babini, Tagliazucchi, Martini, Dei Più, & Gianotti, 2017).As reported in the on line supplementary Tables S1-S8, several tyrosine-and tryptophan-containing peptides were found in the digested milk, which can contribute to the ABTS and hydroxyl radical scavenging activity of the peptidic fractions of milk.
In this study, 43 peptides identified presented ACE inhibition (Table 4).Some identified ACEinhibitory peptides have very low IC50 values and could be the primary contributors to the ACEinhibitory activity of the digested milk (Matsufuji et al., 1994;Nakamura, Yamamoto, Sakai, & Takano, 1995;Kim, Byun, Park, & Shahidi, 2001;Quiros et al., 2007;Kaiser et al., 2016;Tagliazucchi et al., 2016a).Despite the differences in the ACE-inhibitory activity of the digested milk (Table 2) there is no clear species-specific release of ACE-inhibitory peptides.Probably, the diverse activity of the digested milk reflects differences in the amount of released ACE-inhibitory peptides.Three released peptides, namely VPP (identified in the digested cow, goat and sheep milk), IPP (identified in the digested milk of all the studied species) and WL (identified only in digested cow milk) have demonstrated anti-hypertensive activity in humans.In particular, the lactotripeptides VPP and IPP have been shown (at dosages between 5 and 100 mg day −1 ) to decrease the systolic (4.0 mmHg) and diastolic (1.9 mmHg) blood pressure in hypertensive patients and to positively modulate pulse wave velocity in mildly hypertensive subjects (Cicero, Fogacci, & Colletti, 2017).Two recent studies showed that VPP and IPP could be released from cow and goat milk during in vitro digestion at doses, which can elicit physiological effects (Rutella et al., 2016;Tagliazucchi et al., 2017).The α-lactalbumin derived dipeptide WL was found to be bioavailable in human subjects, reducing in vivo ACE activity (Kaiser et al., 2016).One additional peptide (LHLPLP) was found to be able to decrease systolic and diastolic blood pressure in spontaneously hypertensive rats (Quiros et al., 2007).Some other peptides with very low IC50 values (IY, VF and LPP) have been found in plasma of human volunteers after consumption of dairy products (van Platerink, Janssen, Horsten, & Haverkamp, 2006;Foltz et al., 2007).The peptide VY seems to be particularly interesting, behaving as a multifunctional bioactive peptide with high ACE-inhibitory and antioxidant activities (Cheng, Che, Xiong, 2010;Tagliazucchi et al., 2016a).The release of VY was common in milk from the different species studied.VY has been also found in human plasma after consumption of a milk beverage, indicating that this peptide is also released in vivo from cow milk caseins and is bioavailable in humans ( Foltz et al., 2007).
Finally, 20 peptides with previously demonstrated DPP-IV-inhibitory activity were identified in the peptidic fractions of digested milk (Table 5).Cow milk was the best source of DPP-IV-inhibitory peptides (17 out of 20) and the sample with the highest DPP-IV-inhibitory activity (Table 2).Two well-known DPP-IV inhibitors, namely IPI (also known as Diprotin A) and VLP (also known as Diprotin B), were released from cow milk after digestion and could be the primary contributor to the DPP-IV-inhibitory activity of the digested cow milk (Table 5).Diprotin A but not Diprotin B was also found in the peptidic fractions of digested goat and sheep milk.

Conclusion
The present study integrated peptides identified by LC-MS/MS with in vitro bioactivities of milk from four different species (cow, camel, goat and sheep) after the application of the harmonized INFOGEST in vitro gastro-intestinal digestion protocol.Whereas goat milk showed the highest apparent digestibility, sheep milk appeared to be the best source of ACE-inhibitory peptides.
Moreover, cow milk was found to be the best source of DPP-IV-inhibitory peptides and antioxidant peptides and amino acids.Peptidomic analysis showed that goat and sheep milk displayed the highest similarity in peptide sequences identified after in vitro digestion.Most of the released bioactive peptides were in common between two or more species and the peptides with the highest ACE-inhibitory activity had previously demonstrated to be bioavailable in humans.
Although this study lays the basis to distinguish milk from different species in the light of their bioactivities and bioactive peptides released during in vitro digestion, limitations have to be considered.The most important is that this research was conducted analysing one sample of milk from each species.Therefore, to expand the results, studies involving more milk samples are required.Finally, further investigations and in vivo trials are needed to establish which of the observed bioactive peptides have physiological significance.

Figure captions Figure 1 .
Figure captions

Figure 2 .
Figure 2. Evolution of ABTS radical scavenging activity during the in vitro digestion of

Figure 3 .
Figure 3. Venn diagrams of peptides obtained from skimmed cow, camel, goat and sheep

Table 1 .
Chemical composition of skimmed cow, camel, goat and sheep milks.Data are expressed as g 100 g -1 of milk.Values represent means ± standard deviation of triplicate determination; different superscript letters within the same row indicate that the values are significantly different (P<0.05).

Table 2 .
Radical scavenging properties, lipid peroxidation inhibitory activity, and angiotensinconverting enzyme (ACE) and dipeptidyl peptidase IV (DPPIV) inhibitory activities of peptidic fractions (< 3 kDa) obtained from cow, camel, goat and sheep milks after in vitro gastro-intestinal digestion.Values represent means ± standard deviation of triplicate determination; different superscript letters within the same column indicate that the values are significantly different (P<0.05).a % of inhibition was determined using the < 3 kDa fractions of the post-pancreatic sample at a concentration of 1 g L -1 of peptides

Table 3 .
Peptides and amino acids with previously described antioxidant properties identified in the peptidic fractions (< 3 kDa) obtained from cow, camel, goat and sheep milk after in vitro gastro-intestinal digestion.Sample in which the peptide was identified (Co: digested cow milk; Ca: digested camel milk; G: digested goat milk; S: digested sheep milk) a b Precursor protein

Table 4 .
Peptides with previously described angiotensin-converting enzyme (ACE)inhibitory activity identified in the peptidic fractions (< 3 kDa) obtained from cow, camel, goat and sheep milk after in vitro gastro-intestinal digestion.Peptides are listed on the basis of their inhibitory potency.