Article pubs.acs.org/JAFC
Hybrid in Silico/in Vitro Approach for the Identification of Angiotensin I Converting Enzyme Inhibitory Peptides from Parma Dry-Cured Ham Luca Dellafiora,∥ Sara Paolella,∥ Chiara Dall’Asta, Arnaldo Dossena, Pietro Cozzini,* and Gianni Galaverna* Department of Food Science, University of Parma, Parco Area delle Scienze 17/a, 43124 Parma, Italy ABSTRACT: The bioactivity assessment of foodborne peptides is currently a research area of great relevance, and, in particular, several studies are devoted to the antihypertensive effects through the inhibition of angiotensin I converting enzyme (ACE). In the present work, a straightforward workflow to identify inhibitory peptides from food matrices is proposed, which involves a hybrid in vitro/in silico tandem approach. Parma dry-cured ham was chosen as case study. In particular, the advantage of using the hybrid approach to identify active sequences (in comparison to the experimental trials alone) has been pointed out. Specifically, fractions obtained by in vitro gastrointestinal digestion of ham samples of 18 and 24 months of aging have been assessed for ACE inhibition. At the same time, the released peptidomic profiles, which cannot be entirely evaluated by using in vitro assays, have been screened for the inhibition by using an in silico model. Then, to identify novel inhibitory sequences, a series of strong candidates have been synthesized and assessed for their inhibitory activity through in vitro assay. On the one hand, the use of computational simulations appeared to be an effective strategy to find active sequences, as confirmed by in vitro analysis. On the other hand, strong inhibitory sequences were identified for the first time in Parma dry-cured ham (e.g., LGL and SFVTT with IC50 values of 145 and 395 μM, respectively), which is a product of international dietary and economic relevance. Therefore, these findings demonstrate the usefulness of in silico methodologies coupled to in vitro tests for the identification of potentially bioactive peptides, and they give an important contribution to the study of the overall nutritional value of Parma ham. KEYWORDS: bioactive peptides, ACE, molecular modeling, in silico study, dry-cured ham
■
INTRODUCTION Proteins are fundamental components of food; their nutritional value is mainly evaluated in terms of amino acid composition and availability. 1 However, proteins of both plant and animal origin may be also sources of bioactive peptides (BPs), which are protein fragments that, beyond nutritional properties, bring benefits to health of whole organisms at realistic physiological level by exerting a number of biological effects.2−4 In the past decades intensive investigations have been performed for the identification of bioactive sequences in a number of food matrices, opening an intriguing scenario for the reevaluation of the nutritional significance of most foods. Also, the discovery of new active sequences leads to the possibility of introducing specific peptides within dietary supplements or foods (i.e., functional foods) intended to positively affect human health and well-being. In this sense, another promising end point is the opportunity to design nonpharmacological treatments by disposing the ad hoc dietary consumption of naturally rich sources of BPs. Therefore, from a nutrigenomic approach to counteract mild physiological disorders, the discovery of active sequences may lead to “natural” and low-cost treatments, thus avoiding harmful side-effects usually induced by drugs. 5 Within this framework, the BP composition of food has to be actually profiled beforehand and, to this end, the development of novel and high-performing in vitro screening techniques is strongly recommended. Nevertheless, the search through purely experimental investigation may be highly challenging and expensive (vide infra). © 2015 American Chemical Society
Among BPs, angiotensin I converting enzyme (ACE; EC 3.4.15.1) inhibitory peptides are perhaps the most studied, ACE being a target with great pharmacological relevance. 6 The somatic form of ACE is a zinc-dependent carboxypeptidase organized in two distinct catalytic domains that remove dipeptides from the C-terminal end. On the basis of their mechanism of action, ACE inhibitory peptides can be divided into three groups:7 (i) substrate-type peptides, which include peptides that are hydrolyzed by ACE to only slightly active or inactive peptides, thus leading to an “apparent” inhibition; (ii) true inhibition-type peptides, which include peptides having inhibitory potency that is not affected by the pre-incubation with ACE, meaning that they are truly able to inhibit enzymatic reaction; (iii) pro-drug inhibitors, which include peptides that are substrates for ACE or digestive enzymes but that release true inhibitors upon hydrolysis. Several food matrices may encrypt ACE-inhibitory BPs, and some of them have been proposed as paramount sources for BP production (e.g., ref 8). Among these, pork meat is particularly worthy of note inasmuch as it has been proved to be a source of in vivo active peptides. 9 However, due to the complexity and high molecular weight of meat proteins, most active sequences probably are as yet unrevealed and the protein-derived Received: Revised: Accepted: Published: 6366
March 12, 2015 June 23, 2015 June 26, 2015 June 26, 2015 DOI: 10.1021/acs.jafc.5b02303 J. Agric. Food Chem. 2015, 63, 6366−6375
Article
Journal of Agricultural and Food Chemistry
hydrophobic interactions, whereas the sp2 carbonyl oxygen (O) and the neutral flat amino (N1) probes were used to describe the hydrogen bond acceptor and donor capacity of the target, respectively. All images were obtained using the software PyMol version 1.7 (http:// www.pymol.org). Molecular Modeling. The models for both C- and N-domains of ACE were derived from the Protein Data Bank (http://www.rcsb.org) structures having PDB codes 4APH and 4BZS, respectively.27,28 Protein structures and ligands were processed by using the software Sybyl, version 8.1 (www.tripos.com). All atoms were checked for atom- and bond-type assignments. Amino- and carboxyl-terminal groups were set as protonated and deprotonated, respectively. Hydrogen atoms were computationally added to the protein and energy-minimized using the Powell algorithm with a coverage gradient of ≤0.5 kcal/(mol Å) and a maximum of 1500 cycles. Docking Simulations and Rescoring Procedure. The coupling of GOLD, to perform docking simulations, and HINT software, as rescoring function, has been already proved to be effectively able to evaluate the bioactivity of small molecules,16,20,21,29 including peptides.30 The docking simulations of compounds were performed with the GOLD version 5.1 (CCDC; Cambridge, UK; http://www. ccd.cam.ac.uk). All crystallographic waters and ligands were removed, and 25 poses for each compound were generated. No constraints were set up, and the explorable space was defined in a radius of 10 Å from the centroid of the catalytic site. For each GOLD docking search, a maximum number of 100,000 operations were performed on a population of 100 individuals with a selection pressure of 1.1. Operator weights for crossover, mutation, and migration were set to 95, 95, and 10, respectively. The number of islands was set to 5 and the niche to 2. The hydrogen bond distance was set to 2.5 Å and the van der Waals linear cutoff to 4.0. Ligand flexibility options “flip pyramidal N”, “flip amide bonds”, and “flip ring corners” were allowed. Each best-scored pose according to GOLD scoring function was rescored by HINT. Owing to the huge dimension of the pocket, and to speed the spatial search, molecule positioning has been spatially restrained according to crystallographic pose of the inhibitory drug captopril,31 whose pattern of interaction and volume occupancy are maintained also by all of the ligands cocrystallized so far. The software HINT (Hydrophatic INTeraction)32 was used as the rescoring function on the basis of previous studies attesting to the higher reliability of HINT scoring with respect to other scoring functions, as well as its successful use in the search for ligands for other targets and in the estimation of ligand-binding free energies. In more detail, the score provides the evaluation of thermodynamic benefits of protein−ligand interaction, and therefore low/negative scores indicate no appreciable protein−ligand interactions.17,19−21,29,30,33−35 GOLD uses a Lamarckian genetic algorithm, and scores may slightly change from run to run. Therefore, to exclude a noncausative score assignment, we conducted simulations in quintuplicate, and the mean values are reported. In Vitro Experimental Procedures. Samples. Parma dry-cured ham samples with 18 and 24 ripening months were provided by the Experimental Station for Food Preserving Industry (Parma, Italy). Samples of biceps femoris muscle were minced by a common mill (Moulinex, Milano, Italy) and stored at freezing temperature (−22 °C) until analysis. Reagents and Solvents. Doubly distilled water was produced in our laboratory by a Millipore Alpha Q purification system (Waters, Billerica, MA, USA). Pepsin from porcine gastric mucosa, trypsin from porcine pancreas, α-chymotrypsin from bovine pancreas, α-amylase from barley malt (type VIIIa), uric acid, mucin from porcine stomach (type III), glucose, glucuronic acid, glucosamine hydrochloride, bovine serum albumin, pancreatin from porcine pancreas, lipase from porcine pancreas (type II), bovine and ovine bile, sodium dihydrogen phosphate, potassium chloride, sodium hydroxide, sodium phosphate dibasic dodecahydrate, monobasic potassium phosphate, acetonitrile, urea, methanol, trifluoroacetic, triisopropylsilane acid, DL-dithiothreitol, hippuryl-histidyl-leucine (HHL), angiotensin I converting enzyme (ACE; EC 3.4.15.1) from rabbit lung 0.1 U, and glycerol were all purchased from Sigma-Aldrich (St. Louis, MO, USA). Hydrochloric
inhibitory sequences are thus largely unexplored. Although systematic analysis is advisable, the identification and isolation of novel inhibitory peptides in hydrolyzed protein mixtures are challenging tasks: indeed, they are usually based on sequential chromatographic separations of different fractions or components that have to be characterized and tested to identify putative sequences that must then be synthesized and tested for bioactivity. Thus, the development of novel methodologies for the discovery of new active sequences encrypted in food is strongly recommended. Within this framework, the computerdriven screening of peptide sequences may strongly support the discovery of new BPs. Recently a number of in silico approaches have been successfully applied for the discovery of novel bioactive sequences, also toward ACE enzyme (e.g., ref 10). Basically, two main computer-aided frameworks have been applied so far. The former is based on the in silico simulation of enzymatic cleavage or digestion of protein sequences and the evaluation of the release of bioactive sequences (e.g., refs 10 and 11) by using inhibition assay or by checking the release of already known inhibitory peptides (e.g., ref 12). The second is based on the prediction of inhibitory activity of peptides through the use of computational strategies. Notably, most of them use statistical approach to infer activity solely on the basis of the physicochemical properties of peptides (e.g., ref 13). Nevertheless, when the activity requires a certain degree of a protein−ligand interaction, as in the case of peptide-mediated ACE inhibition, structure-based modeling is more advisable.14 Therefore, with the aim to discover novel ACE inhibitory peptides from a previously reported peptidomic profile of digested samples of Parma dry-cured ham,15 we report the benefits achieved by using a hybrid approach involving the integration of in vitro trials and in silico structure-based molecular modeling technique, the reliability of which in predicting the bioactivity of compounds, including peptides, has been previously assessed.16−21 Parma dry-cured ham has been chosen as the case study because it is one of the most valuable brands in Italy with a significant commercialization in Europe, the United States, and Japan.22,23 Even if it is among the most consumed traditional Italian cured-meat products,24 there is a marked lack of data beyond the canonical nutritional values. Thus, the profiling of bioactive compounds, including BPs, is strongly advisable, also considering the efforts to better understand, and rationally improve, the content of health-protective compounds. In particular, dry-cured ham samples have been hydrolyzed by an in vitro simulated gastrointestinal digestion, and then the digested peptide mixtures have been fractionated by preparative LC and the ACE inhibitory activity of each fraction has been evaluated in vitro. In parallel, the priority ranking of peptides in terms of ACE-inhibitory activity has been done on the basis of computational predictions. Finally, in accordance with computational rank and relative occurrence within the most active fractions, some peptides have been synthesized and then assessed for their inhibitory activity by using an in vitro assay.
■
MATERIALS AND METHODS
In Silico Procedures. Pharmachopore Models. The anatomy of the ACE binding sites was investigated by using the Flapsite tool of FLAP software (Fingerprint for Ligand And Protein; http://www. moldiscovery.com),25 and the GRID molecular interaction fields (MIFs) were used to investigate the corresponding pharmacophoric space.26 The DRY probe was used to describe the potential 6367
DOI: 10.1021/acs.jafc.5b02303 J. Agric. Food Chem. 2015, 63, 6366−6375
Article
Journal of Agricultural and Food Chemistry acid (37% v/v), sodium hydroxide, sodium hydrogen carbonate, calcium chloride, ammonium chloride, potassium thiocyanate, and potassium hydroxide were purchased from Carlo Erba (Milan, Italy). Formic acid was purchased from Acros Organics (Geel, Belgium). Sodium chloride was purchased from AnalaR Normapur (Milan, Italy). Sodium sulfate was purchased from Riedel de Haen (Seelze, Germany). Magnesium chloride hexahydrate and boric acid were purchased from Fluka (Sigma-Aldrich, St. Louis, MO, USA). Simulated Gastrointestinal Digestion. Simulated gastrointestinal digestion was recently applied to the study of the effect of dry-cured ham maturation time on the composition of the released peptide fraction:15 the preparation of the relative artificial digestive juices (simulated salivary juice, simulated gastric juice, simulated duodenal juice, and bile) was performed according to the protocol of Versantvoort.36 Briefly, the digestion procedure was as follows: minced biceps femoris muscle (2 g) was dispersed in 3 mL of simulated salivary juice (pH 6.8 ± 0.2) and incubated at 37 °C under agitation with a vortex and a horizontal shaker (230 rpm) for 5 min. After this first step, 6 mL of the simulated gastric juice (pH 1.3 ± 0.2) was added, and the mixture was left to react for 2 h under the same conditions. The last step was the addition of bicarbonate (1 mL, 1 M), simulated duodenal juice (6 mL, pH 8.1 ± 0.2), and bile (3 mL, pH 8.2 ± 0.2), and the mixture was left to react at 37 °C for another 2 h under agitation on a horizontal shaker. The enzyme reaction was stopped by heating at 95 °C for 15 min and the mixture was centrifuged at 3200g for 20 min to separate the supernatant (chyme) and the pellet (residual matrix). The supernatant was subjected to a final cleanup step by a Sep-Pak C18 cartridge (Waters Co., Milford, MA, USA): after the cartridge was conditioned with CH3OH (2 mL) and bidistilled water (3 mL), chyme was applied (2 mL), and after the first eluate had been discarded, peptides were eluted with 2 mL of a solution of H2O/CH3OH (50:50, v/v). The eluate (2 mL) was evaporated to dryness under nitrogen flow in an Eppendorf tube, and the residue was resuspended with 2 mL of aqueous formic acid (0.1%), centrifuged and, then, fractionated in semipreparative HPLC-UV (conditions reported below). Semipreparative LC-UV Fractionation of Digested Samples. The digested mixture was fractionated by semipreparative HPLC-UV (λ = 214nm) by collecting one fraction per minute. The digested samples were separated by a RP column (JUPITER 5 μm, C18, 300 Å, 250 × 10 mm, Phenomenex, Bologna, Italy) using a semipreparative HPLC-UV (1525 binary HPLC pump, Waters, Billerica, MA, USA). The following gradient conditions were applied: eluent A, water with 0.1% formic acid and 0.2% acetonitrile; eluent B, acetonitrile with 0.1% formic acid; gradient, 0−12 min, 100% A; 12−77 min, linear from 100 to 50% A; 77−81 min, 50% A; 81−82 min, linear from 50 to 0% A; 82−90 min, 0% A; 90−91 min, linear from 0 to 100% A; 91−110 min, 100% A. Other parameters were as follows: flow rate, 5 mL/min; analysis time, 110 min; column temperature, room temperature; sample temperature, room temperature; injection volume, 2 mL; fractions collected every minute for a total of 51 fractions. The UV detector (2998 photodiode array, Waters) set at λ = 214 nm. All collected fractions were evaporated to dryness at the rotavapor and dissolved in 2 mL of H2O + 0.1% HCOOH for analyses. UPLC-ESI-MS Analyses. UPLC/ESI-MS analyses were performed with a UPLC/ESI-MS system (UPLC Acquity Waters equipped with a single-quadrupole mass spectrometer, Waters Acquity Ultraperformance). Conditions were as follows: column, RP ACQUITY UPLC BEH 300, C18 (1.7 μm, 2.1 × 150 mm, Waters); gradient elution, eluent A, water with 0.1% formic acid and 0.2% acetonitrile, eluent B, acetonitrile with 0.1% formic acid; gradient, 0−7 min, 100% A; 7−50 min, linear from 100 to 50% A; 50−52.6 min, isocratic, 50% A; 52.6−53 min, linear from 50 to 0% A; 53−58.2 min, isocratic 0% A; 58.2−59 min, linear from 0 to 100% A; 59−72 min, isocratic 100% A. LC parameters: flow rate, 0.2 mL/min; analysis time, 72 min; column temperature, 35 °C; sample temperature, 18 °C; injection volume, 10 μL for fraction samples; 2 μL for the calibration curve of LGL peptide. MS parameters: full scan mode; acquisition time, 7−58.2 min; ionization type, ESI+ (positive ions); scan range, m/z 100−2000;
capillary voltage, 3.2 kV; cone voltage, 30 V; source block temperature, 150 °C; desolvation temperature, 300 °C; cone gas flow, 100 L/h; desolvation gas flow, 650 L/h. Data were acquired and analyzed by MassLynx 4.0 software (Waters Co., Milford, MA, USA). HPLC-ESI-MS/MS Analyses. HPLC/ESI-MS/MS analyses were performed using a HPLC (model Alliance 2695, Waters) equipped with a triple-quadrupole mass spectrometer (model Four Micro, Waters) and a RP column JUPITER 5 μm, C18, 300 Å, 250 × 2 mm i.d. (Phenomenex). Gradient elution was as follows: eluent A, water with 0.1% formic acid and 0.2% acetonitrile; eluent B, acetonitrile with 0.1% formic acid; gradient, 0−12 min, 100% A; 12−77 min, linear from 100 to 50% A; 77−81 min, 50% A; 81−82 min, linear from 50 to 0% A; 82−90 min, 0% A; 90−91 min, linear from 0 to 100% A; 91−110 min, 100% A. Samples (fractions) were first analyzed in full scan mode to identify the characteristic ions and the retention time of the unknown compounds and then in daughters scan modality, using a variable collision energy from 10 to 30 eV. HPLC/ESI-MS/MS parameters were as follows: flow rate, 0.2 mL/min; analysis time, 110 min; column temperature, 35 °C; sample temperature, 23 °C; injection volume, 10 μL for full scan mode and 30 μL for daughter scan mode; acquisition time, 90 min; ionization type, ESI+ (positive ions); scan range, m/z 100−2000; capillary voltage, 3.2 kV; cone voltage, 35 V; source block temperature, 100 °C; desolvation temperature, 200 °C; cone gas flow, 100 L/h; desolvation gas flow, 650 L/h. The peptide sequences were assigned on the basis of the obtained mass spectra and using the Bioinfomatics resource portal ExPasy with the tools FindPept (Swiss Institute of Bioinformatics, Switzerland) and with the web application “Proteomics Toolkit” (Institute for Systems Biology, Seattle, WA, USA). Solid Phase Peptide Synthesis and Purification. Peptides were synthesized using solid phase peptide synthesis according to the Fmoc/tert-butyl strategy on Wang-resin (Wang, CalbiochemNovabiochem, Läufelfingen, Switzerland) using a Syro I Fully Automated Peptide Synthesizer (Biotage, Uppsala, Sweden). Cleavage from the resin was performed using a trifluoroacetic acid (TFA)/ triisopropylsilane (TIS)/H2O/dithiothreitol (DTT) (94:1:2.5:2.5) solution, and peptides were purified by semipreparative RP-HPLCUV (λ = 214 nm). The purity and molecular mass of peptides were determined by using a liquid chromatograph coupled to a mass spectrometer equipped with an electrospray ionization source (UPLC-ESI-MS). Evaluation of ACE Inhibition. The percentage of ACE inhibitory activity for pure peptides and digested fractions was determined by using the methods of Cushman et al 37 and Nakamura et al,38 with some modifications and according to the following equation: I% = [(ACEmax − Bmax) − (ACEmin − Bmin)]/(ACEmax − Bmax) × 100, where ACEmax is the maximum activity of ACE (in the absence of the peptides), ACEmin is the minimal activity of ACE (in the presence of the peptides), Bmax is the control blank of ACE and Bmin is the control blank of sample/digestion blank/pure peptide. The following solutions were prepared: sodium borate buffer (0.1 M, NaBB) with NaCl (300 mM), pH 8.3; potassium phosphate buffer (0.01 M, KPB) with NaCl (500 mM), pH 7; 5 mM hippurylhistidyl-leucine (HHL) in NaBB buffer; and ACE 0.1 U/mL in KPB + 5% glycerol (g/mL). The experiment is carried out at 37 °C in a thermostatic bath (WB-OD 24, Falc, Treviglio (BG), Italy) determining the following parameters: maximum activity of ACE (ACEmax) = 200 μL of HHL + 80 μL of NaBB + 20 μL of ACE; control blank of ACE (Bmax) = 200 μL of HHL + 80 μL of NaBB + 20 μL of KPB; minimal activity of ACE (ACEmin) = 200 μL of HHL + 80 μL of sample (i.e., fraction or pure peptide) + 20 μL of ACE; control blank of sample (i.e., fraction or pure peptide) (Bmin) = 200 μL of HHL + 80 μL of sample (i.e., fraction or pure peptide) + 20 μL of KPB. After 60 min of incubation, the reaction was quenched with 250 μL of HCl (1 N). The analysis was performed by HPLC-UV (Alliance 2695 separation, Waters, Billerica, MA, USA) with dual λ absorbance detector model 2487 (Waters), with the following conditions: column, RP JUPITER C18, 5 μm, 300 Å, 250 × 2 mm (Phenomenex); gradient elution, eluent A, water with 0.1% formic acid and 0.2% acetonitrile; eluent B, acetonitrile with 0.1% formic acid; gradient, 0−10 min, 100% A; 10−22.50 min, linear from 6368
DOI: 10.1021/acs.jafc.5b02303 J. Agric. Food Chem. 2015, 63, 6366−6375
Article
Journal of Agricultural and Food Chemistry
Figure 3. Three-dimensional representation of ACE structures: (A) structures overlapping C-terminal (white) and N-terminal (yellow) domains; (B) cut surface of N-terminal domain. The binding pocket is represented in mesh, and the dark sphere represents Zn ion. Figure 1. ACE-inhibitory activity of fractions collected after an in vitro gastrointestinal digestion of dry-cured hams at 18−24 months of aging time.
LGL at different concentrations (from 1 to 125 μM) containing 4 μL of internal standard (1 mM Phe-Phe) in UPLC-ESI-MS with the analysis conditions reported above.
■
100 to 50% A; 22.50−23.50 min, 50% A; 23.50−30 min, linear from 50 to 100% A; column temperature, 35 °C; sample temperature, room temperature; injection volume, 10 μL for sample fractions and 20 μL for pure peptides; acquisition time, 30 min; flow, 0.2 mL/min; UV detection, λ = 228. Data analysis was performed with Empower software (Waters Co., Milford, MA, USA). The IC50 value is defined as the inhibitor concentration that is able to decrease ACE activity by 50%. To determine IC50, different concentrations of peptides were prepared and their relative ACE inhibitory activity was evaluated. IC50 values were determined by plotting the percentage relative inhibition as a function of concentration of test compound. Peptide Quantification. LGL was quantified by an external calibration curve. The calibration curve was set up by injecting 200 μL of
RESULTS AND DISCUSSION ACE Inhibitory Activity of Digested Fractions. We have recently reported the characterization of the peptide mixture obtained by simulated gastrointestinal digestion of Parma drycured hams:15 To characterize the ACE inhibitory activity of the obtained peptides, the digesta were fractionated by semipreparative HPLC, and the different fractions were tested by in vitro assay, according to common procedures used to identify active sequences.39 By performing a reverse phase HPLC separation with very tight fraction collection, we obtained 51 fractions: results of in vitro test are reported in Figure 1.
Figure 2. Mass spectra of fractions 42, 46, and 47 of samples at 18 months of aging time. 6369
DOI: 10.1021/acs.jafc.5b02303 J. Agric. Food Chem. 2015, 63, 6366−6375
Article
Journal of Agricultural and Food Chemistry
Figure 4. Anatomy of the binding sites: (A) C-terminal domain; (B) N-terminal domain. White, red, and blue contours identify regions sterically and energetically favorable for hydrophobic, H-bond acceptor, and H-bond donor groups, respectively (black regions are due to the back-lighted cut of some contours). The shape of the binding site is represented in mesh.
The inhibitory activity significantly differs among the different fractions, and differences were also observed as a function of aging time. In this respect, indeed, the most active fractions for digested ham sample at 24 months were found to be fractions 6, 14, 46, and 47, whereas for samples at 18 months also fraction 42 showed a relevant activity. To identify the peptides occurring in these fractions, MS and MS/MS analyses have been performed as previously reported.15 The mass spectrum revealed a prevalence of dipeptides and free amino acids (mainly, tyrosine and phenylalanine) in the early eluted fractions (for both aging time). Conversely, longer peptides were found in fractions 42, 46, and 47: the identified sequences are RVAPE, IQLVEEELDRA, DIDDLELT, DIDSPPITAR, LKGADPEDVITGA, and GVVPL, as reported in Figure 2. Nevertheless, many other fractions have shown significant ACE inhibitory activity. It is worth noting that the activity of a specific fraction is determined by two possible contribution: (i) the activity of the occurring peptides or (ii) the total amount of peptides. Indeed, it is possible that a particular fraction shows high activity for the presence of a particularly active sequence in low amount or to the occurrence of a less active peptide in higher amount. Moreover, synergic effect may not be excluded. This mode of investigation is certainly time-consuming, and it is extremely difficult to quantify the peptides both with the total nitrogen evaluation method (Kjieldahl, owing to the low amount in collected fraction) and by chromatographic analysis (lack of authentic standards). Therefore, we have decided to use an innovative computer-aided framework for the research of novel active peptides by using the unbiased computational screening of sequences in comparison to the canonical piecemeal fractioning just applied. In contrast with most other in silico approaches used for the analysis of ACE inhibition, we presented a structure-based procedure, based on the evaluation of protein−ligand interaction, which provides as output the qualitative evaluation of inhibitory activity together with the putative architecture of binding of each ACE−peptide complex. It should be kept in mind that insights concerning the spatial organization of enzyme− inhibitor complex are fundamental for the in-depth understanding of molecular mechanisms of action. Furthermore, they can serve for the rational design/optimization of strong inhibitors. As a general statement, it should be mentioned also that virtual screening can be a fast and cost-effective technique for analyzing huge amounts of sequences. Therefore, the systematic
Figure 5. Comparison between catalytic sites of C-terminal and N-terminal domains. The shape of the pocket is represented in mesh, whereas lining amino acids are represented in sticks. The Zn ion is represented by a dark sphere. Residues of C-terminal and N-terminal domains are colored in white and yellow, respectively. Amino acid substitutions are highlighted by labels.
use for the analysis of peptidomes found in food digesta and/or hydrolyzed mixtures can be a promising straightforward first step toward the total profiling of the BP composition of food. Computational Results. The underlying assumption at the basis of the use of docking-based strategy relies on the concept that the protein−ligand interaction is the sine qua non condition for the inhibitory activity. This means that the protein−ligand interaction is a necessary condition, even if occasionally not sufficient, for the enzymatic inhibition. Consequently, in the case of ACE, peptides able to stably interact with at least one of the two catalytic sites can reduce the conversion of substrates (e.g., angiotensin I in in vivo condition and HHL in in vitro assay), leading to true, substrate-like or pro-drug inhibition mechanisms. For this reason, the evaluation of the binding event at the catalytic site through docking simulations can be a straightforward technique to predict inhibitory potential of peptides. Specifically, to evaluate the enzyme−peptide complex formation, the coupling of docking simulations and rescoring procedures by using the HINT scoring function was chosen, the correlation with the free energy of binding being previously reported. 17 Anatomy of the Pocket. The two catalytic domains originated from tandem gene duplication40 and show the same 3D 6370
DOI: 10.1021/acs.jafc.5b02303 J. Agric. Food Chem. 2015, 63, 6366−6375
Article
Journal of Agricultural and Food Chemistry Table 1. HINT Scores and Expected Activity of Peptides under Analysis
a
peptide sequencea
C-domain
N-domain
expected interaction
RVAPEEHPT VAPEEHPT LAPST FQPS APLNPK PGIAD KLEGDLK YPIEH methylated TEAPLNPK IVAPG MDLER ELV ALM* NSIM* IIAPPER DIR IGGSI MDLE LQDLVDK LTEAPLNPK PSIV SFVTT* DPV DLTDY NWDDMEK NVPI KMEGDLNEM LTL LEGI SYELPDGQ PTVE LGL* IKAKSALA PEILPDGDHDLK SLSTEL LVL VEPEILPDGDHDLK TSLINTK VTV LKGADPEDVITGA DIDSPPITAR GVVPL* INAEL PEILPDGDHDL DQIISANPL LLASIDIDHT TVKDLQHRL IQLVEEELDRA DIDDLELT LFDKPVSPL RMKKNMEQTVK RVAPE INTTLETKQ NAYEESLDQLETLK
−5202 −3113 1046 −494 −396 195 −211 −1238 < −10000 267 1434 503 1885 1001 −2734 −677 846 729 −1342 −5231 688 1481 1398 −5 −2130 400 −3471 1030 169 −2173 428 2503 251 < −10000 −1359 1473 < −10000 1094 1930 < −10000 −3137 1075 −92 1100 395 956 >1100
Figure 6. (A) TIC of LGL in the presence of ACE and HHL; (B) XIC of peptide LGL; (C) XIC of peptides LG/GL.
a huge pocket with similar shape, which crosses the entire protein body (Figure 3). Analyses were focused on catalytic sites thus retracing the mode of action of inhibitory drugs (e.g., captopril and lisinopril).41,42 The regions lining the catalytic site maintain the same organization in both domains, and both pockets share a prevalently hydrophobic environment (Figure 4), albeit they differ for seven amino acid substitutions (Figure 5). Data Collection. The peptide library was based on the peptidomic profile previously characterized after the in vitro physiological digestion of 18 and 24 months aged dry-cured ham.15 However, in this study, solely those sequences that were unequivocally identified were considered, thus discarding those for which the Ile-Leu uncertainty subsisted (same Mw). Moreover, because most dipeptides were already known as active like I(L)F, GI(L), AI(L) (BIOPEP database, http://www.uwm.edu.pl/ biochemia/index.php/pl/biopep), they were excluded throughout. Overall, a total of 54 sequences were analyzed (Table 1). Predicted Activity. All of the computational results obtained for the analyzed sequences are reported in Table 1. A total of 25 peptides were predicted as active, and they ranged from three to five amino acids, with the exception of three peptides (i.e., IKAKSALA, SLSTEL, and TSLINTK). Lead Peptide Identification and in Vitro Assessment of ACE Inhibition. Keeping in mind that the discovery of novel active sequences for future and more detailed analysis is one of the main goals of this work, we decided to synthesize and test for the in vitro ACE inhibition some of the strongest candidates identified by in silico analysis. Because we applied a previously assessed in silico procedure (as aforementioned), which has been followed by experimental confirmations of positive hits, no further validation of the model was needed. Thus, a total of five potentially active peptides were selected by
Peptides selected for experimental trials are denoted by an asterisk (*).
organization with 51% of sequence identity (according to global alignment by using the Needleman−Wunsch algorithm; http:// www.ebi.ac.uk/Tools/psa/emboss_needle). Both domains hold 6371
DOI: 10.1021/acs.jafc.5b02303 J. Agric. Food Chem. 2015, 63, 6366−6375
Article
Journal of Agricultural and Food Chemistry
Figure 7. (A) TIC of mixture NSIM in the presence of ACE and HHL; (B) XIC of peptide NSIM; (C) XIC of peptide IM; (D) fragmentation in source of IM and theoretical fragmentation.
LGL occurred in fractions 32, 31, 33, and 37, respectively. Interestingly, NSIM and SFVTT were found in fractions with no appreciable activity, whereas ALM and LGL were found in fractions with a relatively low ACE-inhibitor activity (Figure 1). Conversely, GVVPL was present in one of the most active fractions (no. 42; Figure 1). Interestingly, to the best of our knowledge, these inhibitory sequences have never been identified previously. Because the GL dipeptide is an already known ACE inhibitor43 (not found in ham digesta), it was assessed whether LGL eventually acted as a pro-drug inhibitor (e.g., peptide converted to true inhibitor after hydrolysis by ACE 44) upon enzymatic cleavage. Thus, the presence of the GL fragment in the inhibitory assay mixture was checked by UPLC-ESI-MS analysis. As reported in Figure 6, the GL fragment was not detected, and therefore a true inhibition mechanism can be proposed for LGL itself. The same was proposed also for ALM, SFVTT, and GVVPL because hydrolysis fragments were not observed (data not shown). Conversely, with regard to NSIM, a substrate mechanism can be proposed because the IM fragment was released in the assay medium (Figure 7). The pro-drug mechanism was excluded due to the low potency observed. As far as the utility of this approach, we can easily affirm that, by the evaluation of the activity of single fractions, the GVVPL peptide would have been identified accordingly, due to the high abundance in one of the most active fractions. However, our results showed that the unbiased analysis of the entire peptidomic profile by using the in silico method allowed the
taking into account the in silico results and their occurrence in active fractions. The inclusion criteria were based (i) on the finding that short peptides may be easily absorbed through the intestinal epithelium,5 thus having a greater physiological significance; (ii) on the relative abundance within the most active fraction of digested mixture, approximately evaluated on the basis of the most intense signals in the mass spectrum;15 and (iii) on the computational results, with the aim to cover a reasonably high range of scores. Therefore, sequences ALM, NSIM, SFVTT, LGL, and GVVPL were synthesized by solid phase peptide synthesis, purified, and then assessed for ACE inhibitory activity. Specifically, LGL showed highest scores for both ACE domains; GVVPL, which was detected in relevant abundance in fraction 42, showed a putative interaction only with the C-domain; and ALM, NSIM, and SFVTT were arbitrarily selected among the remaining peptides with high scores. With respect to this, although a correlation between HINT scores and inhibitory activity cannot be sustained yet, because HINT scores have proved to be correlated with the free energy of binding (as aforementioned), it is reasonable to think that the selection of high-scoring peptides may increase the chances to discover truly active peptides. The inhibitory activity of peptides is expressed as half-maximal inhibitory concentration (IC50), and results of ALM, NSIM, SFVT, LGL, and GVVPL are reported in Table 2. All of the tested peptides showed appreciable inhibitory activity, and the most active sequences were LGL, SFVTT, and GVVPL with IC50 values of 145, 395, and 956 μM, respectively. ALM, NSIM, SFVTT, and 6372
DOI: 10.1021/acs.jafc.5b02303 J. Agric. Food Chem. 2015, 63, 6366−6375
Article
Journal of Agricultural and Food Chemistry
Figure 8. Binding architectures of the most active peptides: (A) LGL within the C-terminal domain; (B) LGL within the N-terminal domain; (C) SFVTT within the C-terminal domain; (D) SFVTT within the N-terminal domain; (E) GVVPL within the C-terminal domain; (F) GVVPL within the N-terminal domain. Amino acids involved in polar contacts and peptides are represented in thin and bold sticks, respectively. The Zn ion is represented in a dark sphere, whereas yellow dotted lines indicate polar contacts. The residues involved in the differential interaction of GVVPL are represented in red sticks, and the hydrophobic−polar mismatch is ringed in red. Because the interaction is predicted as missing, polar contacts between protein and GVVPL are not represented.
predicted owing to charged interferences on valine in position 2. Therefore, it may be hypothesized that the proper substitution with a smaller side chain might increase the activity by gaining interaction with the N-terminal domain. Likewise, with the aim to pursue a domain-specific inhibitor, the loss of the interaction solely with one of the two catalytic sites could be searched. The F391Y mutation can be accounted, accordingly. Quantification of LGL in Ham Samples. LGL is the most active peptide among those considered herein for in vitro assay. Actually, the gastrointestinal stability of inhibitory peptides and the resistance to cleavage by ACE are not always evaluated even if they are essential requirements for seeking in vivo effects. Keeping in mind that all peptides under analysis originated from gastrointestinal digestion of dry-cured ham samples, and considering that di- and tripeptides may be easily adsorbed by intestinal epithelium,5 the true inhibitor LGL might be a sequence of great physiological relevance. Moreover, although effective absorption by intestinal epithelium should be considered, solely on the basis of the IC50 value, LGL might show a congruent range of activity with respect to the abundance per portion. Indeed, to determine the amount of this peptide in dry-cured ham digested samples, a calibration curve was set up plotting the ratio between the peak areas of LGL and FF (as internal standard) and the LGL concentration. In the digested mixtures of 24 and 18 months aged ham samples, the LGL peptide (Mw 301 g/mol) occurs respectively at concentrations of 62 ± 11 and 112 ± 36 μg/g. With regard to the bioaccessibility in respect to the in vitro gastrointestinal digestion
identification of other more active sequences (i.e., LGL and SFVTT), which would have been missed otherwise, being encrypted in low-activity fractions. Theoretically, at the present stage, the in silico method herein proposed is unable to distinguish ACE substrates from true inhibitors or pro-drug peptides because the interaction with the catalytic site is a common requirement. Nevertheless, with only the exception of NSIM, all peptides accounted for in vitro analysis proved to be true inhibitors. Mode of Interaction of Active Peptides. Even if the two domains show some amino acid substitutions at the level of catalytic sites (Figure 5), they did not affect the capability of tested peptides to interact with both domains, with the exception of GVVPL, the binding of which with the N-terminal domain catalytic site was prevented mainly by the F391Y mutation. In particular, the hydroxyl group of Tyr369 within the N-terminal binding site interfered with the side chain of valine in position 2 (from N- to C-terminus), thus preventing the proper positioning of the GVVPL peptide. Therefore, it is proposed that the inhibitory activity of the GVVPL peptide is mediated preferably through the inhibition of the C-domain, whereas the other four peptides can have effects on both catalytic sites. The binding architectures of the three best peptides (LGL, SFVTT, and GVVPL) are reported in Figure 8. Furthermore, it is worth mentioning that the simulation of the binding architecture can be a solid foothold for further enhancing active sequences. For example, with regard to GVVPL the lack of interaction with the N-terminal domain was 6373
DOI: 10.1021/acs.jafc.5b02303 J. Agric. Food Chem. 2015, 63, 6366−6375
Article
Journal of Agricultural and Food Chemistry model, 829 μg of LGL is needed in to reach a concentration able to decrease ACE activity by 50% (which is contained in 13.4 and 7.4 g of 24 and 18 months aged ham, respectively). Furthermore, such a concentration is overcome by 7.5- and 13.5-fold when 24 and 18 months aged ham portions (100 g) are consumed. Therefore, one can hypothesize that physiologically relevant concentrations might be reached after dietary consumption of Parma ham, at least in the intestinal lumen. Accordingly, the high abundance in ham samples, together with the high activity observed in in vitro ACE inhibition, concur to elect LGL as a strong lead compound and also strongly support its inclusion in future and more detailed experimental investigations. Taken as a whole, our findings are an important step forward toward the in-depth and more informed characterization of such a relevant food product beyond its wellknown nutritional properties.
■
components in blood pressure reduction. Compr. Rev. Food Sci. Food Saf. 2014, 13, 114−134. (8) Yang, Y.; Marczak, E. D.; Yokoo, M.; Usui, H.; Yoshikawa, M. Isolation and antihypertensive effect of angiotensin I-converting enzyme (ACE) inhibitory peptides from spinach Rubisco. J. Agric. Food Chem. 2003, 51, 4897−4902. (9) Nakashima, Y.; Arihara, K.; Sasaki, A.; Mio, H.; Ishikawa, S.; Itoh, M. Antihypertensive activities of peptides derived from porcine skeletal muscle myosin in spontaneously hypertensive rats. J. Food Sci. 2002, 67, 434−437. (10) Lafarga, T.; O’Connor, P.; Hayes, M. Identification of novel dipeptidyl peptidase-IV and angiotensin-I-converting enzyme inhibitory peptides from meat proteins using in silico analysis. Peptides 2014, 59, 53−62. (11) Asoodeh, A.; Haghighi, L.; Chamani, J.; Ansari-Ogholbeyk, M. A.; Mojallal-Tabatabaei, Z.; Lagzian, M. Potential angiotensin I converting enzyme inhibitory peptides from gluten hydrolysate: biochemical characterization and molecular docking study. J. Cereal Sci. 2014, 60, 92−98. (12) Vermeirssen, V.; van der Bent, A.; Van Camp, J.; van Amerongen, A.; Verstraete, W. A quantitative in silico analysis calculates the angiotensin I converting enzyme (ACE) inhibitory activity in pea and whey protein digests. Biochimie 2004, 86, 231−239. (13) Sagardia, I.; Roa-Ureta, R. H.; Bald, C. A new QSAR model, for angiotensin I-converting enzyme inhibitory oligopeptides. Food Chem. 2013, 136, 1370−1376. (14) Gleeson, M. P.; Modi, S.; Bender, A.; Robinson, R. L.; Kirchmair, J.; Promkatkaew, M.; Hannongbua, S.; Glen, R. C. The challenges involved in modeling toxicity data in silico: a review. Curr. Drug Metab. 2012, 18, 1266−1291. (15) Paolella, S.; Falavigna, C.; Faccini, A.; Virgili, R.; Sforza, S.; Dall’Asta, C.; Dossena, A.; Galaverna, G. Effect of dry-cured ham maturation time on simulated gastrointestinal digestion: characterization of the released peptide fraction. Food Res. Int. 2015, 67, 136− 144. (16) Amadasi, A.; Mozzarelli, A.; Meda, C.; Maggi, A.; Cozzini, P. Identification of xenoestrogens in food additives by an integrated in silico and in vitro approach. Chem. Res. Toxicol. 2009, 22, 52−63. (17) Cozzini, P.; Fornabaio, M.; Marabotti, A.; Abraham, D. J.; Kellogg, G. E.; Mozzarelli, A. Simple, intuitive calculations of free energy of binding for protein-ligand complexes. 1. Models without explicit constrained water. J. Med. Chem. 2002, 45, 2469−2483. (18) Delfosse, V.; Grimaldi, M.; Cavaillès, V.; Balaguer, P.; Bourguet, W. Structural and functional profiling of environmental ligands for estrogen receptors. Environ. Health Perspect. 2014, 122, 1306−1313. (19) Dellafiora, L.; Dall’Asta, C.; Cruciani, G.; Galaverna, G.; Cozzini, P. Molecular Modelling approach to evaluate poisoning of topoisomerase I by alternariol derivatives. Food Chem. 2015, DOI: 10.1016/ j.foodchem.2015.02.083. (20) Dellafiora, L.; Mena, P.; Cozzini, P.; Brighenti, F.; Del Rio, D. Modelling the possible bioactivity of ellagitannin-derived metabolites. In silico tools to evaluate their potential xenoestrogenic behavior. Food Funct. 2013, 4, 1442−1451. (21) Dellafiora, L.; Mena, P.; Del Rio, D.; Cozzini, P. Modeling the effect of phase II conjugations on topoisomerase I poisoning: pilot study with luteolin and quercetin. J. Agric. Food Chem. 2014, 62, 5881−5886. (22) Prosciutto di Parma, Consortium; http://www. prosciuttodiparma.com/en_UK/consortium/economic-figures. (23) Cannata, S.; Ratti, S.; Meteau, K.; Mourot, J.; Baldini, P.; Corino, C. Evaluation of different types of dry-cured ham by Italian and French consumers. Meat Sci. 2010, 84, 601−606. (24) Lucarini, M.; Saccani, G.; D’Evoli, L.; Tufi, S.; Aguzzi, A.; Gabrielli, P.; Marletta, L.; Lombardi-Boccia, G. Micronutrients in Italian ham: a survey of traditional products. Food Chem. 2013, 140, 837−842. (25) Baroni, M.; Cruciani, G.; Sciabola, S.; Perruccio, F.; Mason, J. S. A common reference framework for analyzing/comparing proteins and
AUTHOR INFORMATION
Corresponding Authors
*(P.C.) Phone: ++39-0521-905669. E-mail: pietro.cozzini@ unipr.it. *(G.G.) Phone: ++39-0521-906270. E-mail: gianni.galaverna@ unipr.it. Author Contributions ∥
L.D. and S.P. contributed equally to the work.
Funding
This work was supported by the project “Hepiget-Advanced research in genomics and processing technologies for the Italian heavy pig production chain” financed by the AGER-Agroalimentare e Ricerca foundation (Italy). Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS We acknowledge Prof. Glen E. Kellogg (Virginia Commonwealth University, Richmond, VA, USA; USA) for the HINT scoring function, Prof. Gabriele Cruciani (University of Perugia, Perugia, Italy) for kindly supplying the FLAP software, Dario Manfredi for programming support, and Dr. Roberta Virgili (Experimental Station for Food Preserving Industry, Parma, Italy) for the ham samples.
■
REFERENCES
(1) Puchalska, P.; Marina Alegre, M. L.; García López, M. C. Isolation and characterization of peptides with antihypertensive activity in foodstuffs. Crit. Rev. Food Sci. Nutr. 2015, 55, 521−551. (2) Harnedy, P. A.; FitzGerald, R. J. Bioactive peptides from marine processing waste and shellfish: a review. J. Funct. Foods 2012, 4, 6−24. (3) Hartmann, R.; Meisel, H. Food-derived peptides with biological activity: from research to food applications. Curr. Opin. Biotechnol. 2007, 18, 163−169. (4) Kim, S.; Wijesekara, I. Development and biological activities of marine-derived bioactive peptides: a review. J. Funct. Foods 2010, 2, 1− 9. (5) De Leo, F.; Panarese, S.; Gallerani, R.; Ceci, L. R. Angiotensin converting enzyme (ACE) inhibitory peptides: production and implementation of functional food. Curr. Pharm. Des. 2009, 15, 3622−3643. (6) Anthony, C. S.; Masuyer, G.; Sturrock, E. D.; Acharya, K. R. Structure based drug design of angiotensin-I converting enzyme inhibitors. Curr. Med. Chem. 2012, 19, 845−855. (7) Iwaniak, A.; Minkiewicz, P.; Darewicz, M. Food-originating ACE inhibitors, including antihypertensive peptides, as preventive food 6374
DOI: 10.1021/acs.jafc.5b02303 J. Agric. Food Chem. 2015, 63, 6366−6375
Article
Journal of Agricultural and Food Chemistry ligands. Fingerprints for Ligands and Proteins (FLAP): theory and application. J. Chem. Inf. Model. 2007, 47, 279−294. (26) Carosati, E.; Sciabola, S.; Cruciani, G. Hydrogen bonding interactions of covalently bonded fluorine atoms: from crystallographic data to a new angular function in the GRID force field. J. Med. Chem. 2004, 47, 5114−5125. (27) Kramer, G. J.; Mohd, A.; Schwager, S. L.; Masuyer, G.; Acharya, K. R.; Sturrock, E. D.; Bachmann, B. O. Interkingdom pharmacology of angiotensin-I converting enzyme inhibitor phosphonates produced by actinomycetes. ACS Med. Chem. Lett. 2014, 5, 346−351. (28) Masuyer, G.; Schwager, S. L.; Sturrock, E. D.; Isaac, R. E.; Acharya, K. R. Molecular recognition and regulation of human angiotensin-I converting enzyme (ACE) activity by natural inhibitory peptides. Sci. Rep. 2012, 2, 717. (29) Cozzini, P.; Dellafiora, L. In silico approach to evaluate molecular interaction between mycotoxins and the estrogen receptors ligand binding domain: a case study on zearalenone and its metabolites. Toxicol. Lett. 2012, 214, 81−85. (30) Salsi, E.; Bayden, A. S.; Spyrakis, F.; Amadasi, A.; Campanini, B.; Bettati, S.; Dodatko, T.; Cozzini, P.; Kellogg, G. E.; Cook, P. F.; Roderick, S. L.; Mozzarelli, A. Design of O-acetylserine sulfhydrylase inhibitors by mimicking nature. J. Med. Chem. 2010, 53, 345−356. (31) Akif, M.; Masuyer, G.; Schwager, S. L.; Bhuyan, B. J.; Mugesh, G.; Isaac, R. E.; Sturrock, E. D.; Acharya, K. R. Structural characterization of angiotensin I-converting enzyme in complex with a selenium analogue of captopril. FEBS J. 2011, 278, 3644−3650. (32) Kellogg, E. G.; Abraham, D. J. Hydrophobicity: is LogP(o/w) more than the sum of its parts? Eur. J. Med. Chem. 2000, 37, 651−661. (33) Fornabaio, M.; Cozzini, P.; Mozzarelli, A.; Abraham, D. J.; Kellogg, G. E. Simple, intuitive calculations of free energy of binding for protein-ligand complexes. 2. Computational titration and pH effects in molecular models of neuraminidase-inhibitor complexes. J. Med. Chem. 2003, 46, 4487−4500. (34) Fornabaio, M.; Spirakis, F.; Mozzarelli, A.; Cozzini, P.; Abraham, D. J.; Kellogg, G. E. Simple, intuitive calculations of free energy of binding for protein-ligand complexes. 3. The free energy contribution of structural water molecules in HIV-1 protease complexes. J. Med. Chem. 2004, 47, 4507−4516. (35) Marabotti, A.; Spyrakis, F.; Facchiano, A.; Cozzini, P.; Alberti, S.; Kellogg, G. E.; Mozzarelli, A. Energy-based prediction of amino acid-nucleotide base recognition. J. Comput. Chem. 2008, 29, 1955− 1969. (36) Versantvoort, C. H.; Oomen, A. G.; Van de Kamp, E.; Rompelberg, C. J.; Sips, A. J. Applicability of an in vitro digestion model in assessing the bioaccessibility of mycotoxins from food. Food Chem. Toxicol. 2005, 43, 31−40. (37) Cushman, D. W.; Cheung, H. S. Spectrophotometric assay properties of the angiotensin-converting enzyme of rabbit lung. Biochem. Pharmacol. 1971, 20, 1637−1648. (38) Nakamura, Y.; Yamamoto, N.; Sakai, K.; Okubo, A.; Yamazaki, S.; Takano, T. Purification and characterization of angiotensin Iconverting enzyme inhibitors from sour milk. J. Dairy Sci. 1995, 78, 777−783. (39) Escudero, E.; Mora, L.; Fraser, P. D.; Aristoy, M. C.; Arihara, K.; Toldrá, F. Purification and Identification of antihypertensive peptides in Spanish dry-cured ham. J. Proteomics 2013, 78, 499−507. (40) Coates, D.; Isaac, R. E.; Cotton, J.; Siviter, R.; Williams, T. A.; Shirras, A.; Corvol, P.; Dive, V. Functional conservation of the active sites of human and Drosophila angiotensin I-converting enzyme. Biochemistry 2000, 39, 8963−8969. (41) Corradi, H. R.; Schwager, S. L.; Nchinda, A. T.; Sturrock, E. D.; Acharya, K. R. Crystal structure of the N domain of human somatic angiotensin I-converting enzyme provides a structural basis for domain-specific inhibitor design. J. Mol. Biol. 2006, 357, 964−974. (42) Yates, C. J.; Masuyer, G.; Schwager, S. L.; Akif, M.; Sturrock, E. D.; Acharya, K. R. Molecular and thermodynamic mechanisms of the chloride-dependent human angiotensin-I-converting enzyme (ACE). J. Biol. Chem. 2014, 289, 1798−1814.
(43) Cheung, H. S.; Wang, F. L.; Ondetti, M. A.; Sabo, E. F.; Cushman, D. W. Binding of peptide substrates and inhibitors of angiotensin-converting enzyme. Importance of the COOH-terminal dipeptide sequence. J. Biol. Chem. 1980, 255, 401−407. (44) Fujita, H.; Yokoyama, K.; Yoshikawa, M. Classification and antihypertensive activity of angiotensin I-converting enzyme inhibitory peptides derived from food proteins. J. Food Sci. 2000, 65, 564−569.
6375
DOI: 10.1021/acs.jafc.5b02303 J. Agric. Food Chem. 2015, 63, 6366−6375
