In silico investigation of action mechanism offour novel angiotensin-I converting enzymeinhibitory peptides modified with Trp

Cheng-liang Xie a,d, Se-young Choung b, Guang-ping Cao c,Keun Woo Lee c, Yeung Joon Choi a,*a Department of Seafood Science and Technology/Institute of Marine Industry, Gyeongsang National University,Gyeongnam 650-160, South Koreab Department of Preventive Pharmacy & Toxicology, College of Pharmacy, Kyung Hee University, Seoul 130-701,South Koreac Department of Biochemistry, Division of Applied Life Science (BK21 Plus Program), Systems and SyntheticAgrobiotech Center (SSAC), Plant Molecular Biology and Biotechnology Research Center (PMBBRC), ResearchInstitute of Natural Science(RINS), Gyeongsang National University, Gyeongnam 660-701, South Koread Department of Anatomy and Convergence Medical Science, Institute of Health Sciences, School of Medicine,Gyeongsang National University, Gyeongnam 660-751, South Korea

A R T I C L E I N F O

Article history:

Received 25 January 2015

Received in revised form 25 May

2015

Accepted 2 June 2015

Available online

A B S T R A C T

To potentiate activities of angiotensin-I converting enzyme (ACE) inhibitory peptides VK,

YA, KY and TAY from oyster hydrolysate, four novel tripeptides VKW, YAW, KYW and TAW

were designed by modification with Trp at the C-terminus. The tripeptides were synthe-

sized, and their 50% ACE inhibitory concentrations (IC50) were measured to be 0.020, 0.49,

0.43 and 0.63 µM, respectively. Their activities increased 27–1450 times compared to their

corresponding original peptides. All peptides did not show toxicity on the Chang cell. Mo-

lecular docking and molecular dynamics simulation showed that modification with Trp can

enhance the stability of ACE/derived peptide complexes by increasing binding affinity and

interaction sites with important amino acid residues. Interactions between peptides with

amino acid residues of ACE GLN281, ALA354, GLU411, PHE457 and LYS511 played an impor-

tant role in activity improvement of derived peptides. Modification with Trp at the C-terminus

was an effective way for designing novel ACE inhibitory peptides.

© 2015 Elsevier Ltd. All rights reserved.

Keywords:

Angiotensin-I converting enzyme

Peptide

Molecular docking

Molecular dynamics simulation

1. Introduction

Angiotensin-I converting enzyme (ACE, EC 3.4.15.1) is a keytherapeutic target for combating hypertension and related car-diovascular diseases (Zhou, Du, Ji, & Feng, 2012). For this reason,ACE inhibitors become one of the most important therapeutic

agents available for the treatment of hypertension, heart failure,myocardial infraction, and diabetic nephropathy (Acharya,Sturrock, Riordan, & Ehlers, 2003; Natesh, Schwager, Evans,Sturrock, & Acharya, 2004). Bioactive peptides derived from foodprotein are increasingly becoming important as starting pointsfor the development of functional ingredients and drug-related compounds. Among various bioactive peptides, ACE

* Corresponding author. Department of Seafood Science and Technology/Institute of Marine Industry, Gyeongsang National University,Gyeongnam 650-160, South Korea. Tel.: +82 55 772 9143; fax: +82 55 772 9149.

E-mail address: yjchoi@gnu.ac.kr (Y.J. Choi).http://dx.doi.org/10.1016/j.jff.2015.06.0161756-4646/© 2015 Elsevier Ltd. All rights reserved.

J o u rna l o f Func t i ona l F ood s 1 7 ( 2 0 1 5 ) 6 3 2 – 6 3 9

Available online at www.sciencedirect.com

ScienceDirect

journal homepage: www.elsevier.com/ locate / j ff

inhibitory peptides have been studied extensively and a largenumber of ACE inhibitory peptides have been purified and iden-tified from various food sources such as cereals (De Leo,Panarese, Gallerani, & Ceci, 2009; Shamloo, Eck, & Beta, 2015),egg (Majumder et al., 2015), milk (Zhou et al., 2012), casein(Holder et al., 2013; Yamada et al., 2013), porcine muscle(Castellano, Aristoy, Sentandreu, & Vignolo, 2013) and marinesource such as oyster and lizard fish (Wu, Feng, Lan, Xu, & Liao,2015; Xie, Kim, Ha, Choung, & Choi, 2014).

To analyze and interpret the action mechanism of novel ACEinhibitory peptides, in silico approaches will be a strategy takinginto account the large amount of theoretical possible pep-tides, i.e., 400 dipeptides, 8000 tripeptides, 160 000 tetrapeptides,etc. Quantitative structure–activity relationship (QSAR) modelshave been applied to study the relationships between struc-tures and ACE inhibitory activities of peptides (Wu, Aluko, &Nakai, 2006a, 2006b). Wu et al. (2006b) reported that theC-terminal residue of ACE inhibitory peptides plays a predomi-nated role in competitive binding to the active site of ACE. Themost favorable residues for tripeptide at the C-terminus arearomatic or hydrophobic amino acids (Wu et al., 2006a). Prolineis well documented as the most favorable amino acid at theC-terminus and most commercially existing inhibitors bear thisresidue (Wu et al., 2006a). However, there are few studies aboutthe effect of hydrophobic amino acid Trp at the C-terminus onthe ACE inhibitory activity although it has been reported thatACE inhibitory peptides with Trp at the C-terminus show highACE inhibitory activities (Seki, Osajima, Matsufuji, Matsui, &Osajima, 1995; Wu et al., 2006a; Yamada, Matoba, Usui, Onishi,& Yoshikawa, 2002). The crystal structure of the human ACEand inhibitor complex has been elucidated (Natesh, Schwager,Sturrock, & Acharya, 2003), which provides a platform for theaction mechanism analysis of ACE inhibitory peptides by mo-lecular docking and molecular dynamics (MD) simulation. Ithas been reported that the combination of fast moleculardocking with accurate but time-consuming MD techniques hasbeen successfully applied in the discovery of new potential in-hibitors and study on interaction mechanism between inhibitorypeptides and ACE in molecular level (Alonso, Bliznyuk, & Gready,2006; Zhou et al., 2012).

In our previous studies, four ACE inhibitory peptides TAY,VK, KY and YA have been purified and identified from oysterhydrolysate (Xie et al., 2014). Moreover, tripeptides are ab-sorbed easily compared with the long peptides (Wu et al., 2006b).Four novel tripeptides VKW, KYW, YAW and TAW were de-signed by modification with Trp at the C-terminus of peptides.

The objectives of this study were to synthesize the four novelACE inhibitory tripeptides VKW, KYW, YAW and TAW, to de-termine their ACE inhibitory activity and cell toxicities, and toinvestigate the action mechanism between tripeptides and ACEin molecular level by molecule docking and MD simulation.

2. Materials and methods

2.1. Materials

ACE (5 units from rabbit lung), hippuryl-histidyl-leucine (HHL)and captopril were purchased from Sigma-Aldrich Co. (St. Louis,MO, USA). Normal human hepatocyte Chang cell line (ATCC

CCL-13) was obtained from the Korean Cell Line Bank (KCLB,Seoul, Korea). Fetal bovine serum (FBS), modified Eagle’s medium(MEM), trypsin–ethylenediaminetetraacetic acid (EDTA),phosphated-buffered saline–EDTA, penicillin and streptomycinwere purchased from GIOCO BRL (Grand Island, NY, USA). Eightpeptides (VK, VKW, TAY, TAW, YA, YAW, KY and KYW) were syn-thesized by the standard Fmoc-based solid-phase method atGL Biochem Ltd. (Shanghai, China). The purities of the pep-tides were estimated on Agilent 1200 HPLC system with a UVdetector (Agilent Technologies, Waldbronn, Germany) and MSdata were collected with a Shimadzu LCMS-2010 mass spec-trometry (MS) (Shimadzu Co., Kyoto, Japan) (Appendix S1).Acetonitrile was of high performance liquid chromatography(HPLC) grade. Other products were used as reagent grade. AWatchers 120 ODS-AP (5 µm, 4.6 × 250 mm) C18 column waspurchased from Daiso Co. (Tokyo, Japan).

2.2. The property of the peptides

The 3D geometrical structures of peptides were prepared andtheir LogP values, which are the Log of the (1-octanol/water)partition coefficient, were calculated using Hyperchem 8.0 soft-ware (Hypercube, Gainesville, FL, USA).

2.3. ACE inhibitory activity assay

ACE inhibitory activity was measured according to the methodof Wu, Aluko, and Muir (2002) with slight modifications. Samplepeptide (20 µL) was mixed with 225 µL of ACE solution(0.25 units/mL) and pre-incubated at 37 °C for 10 min beforeadding 50 µL of HHL (2.5 mg/mL of 0.1 M borate buffer, pH 8.3containing 0.3 M NaCl). The mixture was incubated for 30 minat 37 °C. The reaction was stopped by adding 75 µL of 1 M HCl,and the mixture was centrifuged at 10 000 rpm (Micro 17TR,Hanil Sci. Industry Co., Inchun, Korea). The supernatant con-taining hippuric acid (HA) was determined by RP HPLC(Shimadzu Systems, Kyoto, Japan) on Watchers C18 (5 µm,4.6 × 250 mm). The IC50 value was calculated as the inhibitorconcentration required to inhibit 50% of ACE activity. Thepercent inhibition of enzyme activity was calculated as follows:

Inhibitionactivity HA HA HAcontrol sample control%( ) = −( ) × 100

2.4. Cell toxicity of the peptides

Chang cells were cultured in MEM medium containing 10% fetalbovine serum. Ninety-six well plates containing 1 × 104 cellsper well were incubated at 37 °C under 95% humidity and 5%CO2. Twenty-four hours later, the medium was replaced withthe medium containing the peptides at the final concentra-tions of 1, 5, 10, and 20 µg/mL and incubated under the sameconditions for an additional 24 h. Cell growth was assessed withthe CellTiter 96 Aqueous One Solution Cell Proliferation Assaykit (Promega, Madison, WI, USA) and absorbance was mea-sured at 490 nm with a microplate reader (Perkin Elmer 1420,VICTOR X Multilabel Plate readers, Waltham, MA, USA). Cell vi-ability was calculated by comparing with the absorbance of theuntreated group.

633J o u rna l o f Func t i ona l F ood s 1 7 ( 2 0 1 5 ) 6 3 2 – 6 3 9

2.5. Molecular docking

The crystal structure of the human testicular ACE/lisinoprilcomplex was selected as working targets, which was obtainedfrom the Protein Data Bank (PDB: 1O86) (http://www.rcsb.org).The PDB file was visualized and processed using Sybyl soft-ware package (Tripos International, St. Louis, MO, USA) byremoving all substrates, chloride ions and water molecular.Thenthe protein structure was pre-analyzed and prepared for thedocking runs using the biopolymer structure preparation toolwith default settings as implemented in the Sybyl programpackage. The 3D structures of peptides were prepared andenergy minimization was performed with the Tripos force fieldusing the Powell conjugate gradient optimization algorithm witha convergence criterion of 0.05 kcal/mol. The Surflex-Dockprogram was selected for docking studies. During the dockingprocess, the ligand (lisinopril) from the ACE–lisinopril complexwas used as a reference molecule and the top 20 conforma-tions were generated and ranked the conformations for eachpeptide based on Total Scores of Surflex-Dock and the otherthree kinds of docking score D_Score, G_score and Chemscorewere calculated together. The conformations of ranked No. 1for each peptide were selected and the interaction modelbetween the ACE residues and peptides was analyzed by Dis-covery Studio (Version 3.5, Accelrys, San Diego, CA, USA).

2.6. MD simulation

In order to improve interactions and enhance complementar-ity between ACE and peptides, the ACE/peptide complexesobtained by molecular docking were refined by a 10 ns MDsimulation. MD simulations were performed with the Gromacs4.5.3 package using AMBER03 force field on a high perfor-mance Linux cluster computer (Lindahl, Hess, & van der Spoel,2001). Topology files of the inhibitors were generated usingACPYPE (AnteChamber Python Parser interface) (Silva &Vranken, 2012). The structure was solvated in a dodecahe-dron box with a length of 1 nm and the TIP3P water model wasgenerated to perform the simulations in an aqueous environ-ment (Berendsen, Postma, van Gunsteren, & Hermans, 1981;Jorgensen, Chandrasekhar, Madura, Impey, & Klein, 1983). TenNa+ counter ions were added by replacing water molecules toensure the overall charge neutrality of the simulated system.The systems were subjected to a step by step descent energyminimization process until a tolerance of 1000 kJ/mol to avoidhigh energy interactions and steric clashes. The energy mini-mized system was treated for 100 ps in an equilibration run.A constant temperature and pressure of 300 K and 1 bar wereachieved with the V-rescale thermostat and Parrinello–Rahmanbarostat (Bussi, Donadio, & Parrinello, 2007; Parrinello &Rahman, 1981). The particle mesh Ewald (PME) method wasapplied to accurately determine the long-range electrostaticinteractions (Essmann et al., 1995) with an interpolation orderof 6 and a grid spacing of 0.12 nm. The cutoff for van der Waalsinteractions was 1.2 nm, and hydrogen bonding interactionswere monitored using 0.35 nm as the donor–acceptor dis-tance cutoff and 60° as the hydrogen-donor–acceptor anglecutoff (Zhou et al., 2012). Bonds between heavy metals and cor-responding hydrogen atoms were constrained to theirequilibrium bond lengths using the LINCS21 algorithm (Hess,

Bekker, Berendsen, & Fraaije, 1997). The time step for the simu-lations was set to 2 fs. During the production phase, thecoordinate data were written to the file every ps.

2.7. Statistical analysis

Data were expressed as the mean with standard deviation oftriplicate determinations. Analysis of variance was carried outby the Tukey HSD test using the JMP 10 package (SAS Insti-tute, Cary, NC, USA).

3. Results and discussion

3.1. The properties of the peptides

The derived peptides were synthesized by solid-phase syn-thesis method. Solid-phase peptide synthesis can be fulfilledby commercial automated instruments and is widely used inthe field of peptide synthesis (Fields, Lauer-Fields, Liu, & Barany,2002).The purities of all synthesized peptides were higher than95% by an HPLC analysis and these peptides were confirmedby comparing the molecular weight from the calculated valuesand the measured values (Table 1). The LogP of the derivedtripeptides increased to the different extents compared withtheir corresponding original peptides (Table 1). LogP is the mostfrequently used physicochemical property to predict cellularpermeability, and an increase in LogP reflects an increase inlipophilicity and often corresponds to an increase in cell per-meability and bioavailability in vivo (Kulkarni, Han, & Hopfinger,2002).

3.2. ACE inhibitory activity

The IC50 values of ACE inhibitory activities of the four derivedtripeptides VKW, YAW, KYW and TAW were 0.020, 0.49, 0.43 and0.63 µM, respectively, and their activities increased 27–1450 timescompared to their corresponding original peptides (Table 1).Especially, the ACE inhibitory activity of VKW was higher 1450times than that of VK.VKW was about 1/180 potency of captopril(IC50 = 0.11 nM) and comparable with potency of the venom

Table 1 – Physicochemical properties and ACE inhibitoryactivity of peptides.

Peptides Purity(%)

Molecular mass(Da)

LogP IC50

(µM)Relativeactivity

Calculated [M + H]+*

VK 95.92 245.32 246.30 −0.43 29 1VKW 95.67 431.53 432.15 0.49 0.020 1450YA 99.74 252.27 253.20 0.33 94 1YAW 95.95 438.48 438.96 1.25 0.49 192KY 95.86 309.37 310.08 0.10 52 1KYW 98.44 495.57 496.15 1.02 0.43 121TAY 99.62 353.38 354.25 −0.58 17 1TAW 95.98 376.41 376.72 −0.52 0.63 27

* Measured values by mass spectrometer; LogP, Log of the(1-octanol/water) partition coefficient calculated by Hyperchem 8.0software.

634 J o u rna l o f Func t i ona l F ood s 1 7 ( 2 0 1 5 ) 6 3 2 – 6 3 9

peptide (Cushman, Cheung, Sabo, & Ondetti, 1977). The fournovel tripeptides especially VKW have a potential to be appliedas drug lead compounds or functional ingredients.

3.3. In vitro cell toxicity

Cell toxicity of the peptides was confirmed by the Chang cellline.The cell viability ranged from 92.4 ± 6.6 to 117.3 ± 9.2% aftertreatment with the synthetic peptides and no adverse effectswere observed compared with untreated group (p < 0.05,Supplementary Table S1). The results were accorded with theeffect of ACE inhibitory peptides on HepG2 cell line in our pre-vious study (Xie et al., 2014).

3.4. Molecular docking

Molecular docking has become a major computational method,which can predict ligand–acceptor interaction and rank com-pounds by estimating the binding affinities of the compoundand receptor complexes (Kellenberger, Rodrigo, Muller, &Rognan, 2004). In order to study the interaction mechanismbetween peptides and ACE, all peptides were docked into theactive sites of ACE by Surflex-Dock module in Sybyl. It has beenreported that Surflex-Dock can accurately discriminate knowninhibitor of an enzyme (thymidine kinase) from randomlychosen molecules and successfully rank known inhibitors ina virtual screening experiment (Kellenberger et al., 2004).Because unique scoring was difficult to generate reliable results,four different docking scores Total Score, D_Score, G_Score andChemscore in Sybyl were used to estimate the binding affin-ity of the ACE/peptide complexes in the study (Table 2). Thefour algorithms of docking score are different, among whichD_Score and G_Score are based on force field that compute theenthalpy gain or loss upon binding, while Total Score andChemscore use empirical terms to estimate free energy (Xing,Hodgkin, Liu, & Sedlock, 2004). The absolute values of fourdocking scores increased for VKW (Total Score, +0.5; D_Score,+60.4; G_Score, +93.6 and Chemscore, +11.0) and YAW (TotalScore, +4.7; D_Score, +50.7; G_Score, +64.9 and Chemscore, +14.8)compared with that of VK and YA. While the absolute valuesof Total Score among four docking scores decreased slightlyfor KYW (−0.1) and TAW (−0.7) compared with that of KY andTAY, which can explain why the increasing degree of ACE in-hibitory activity of KY and TAY was smaller than that of VKand YA after modified by Trp (Table 1). As a whole, the binding

affinities of ACE/peptide complexes increased after modifica-tion with Trp, which were accorded with the experimentalresults (Table 1).

3.5. MD simulation

Molecular docking lacks flexibility of the protein and does notpermit to adjust its conformation for ligand binding, while MDsimulation can treat both ligand and protein in a flexible wayfor an induced fit of the receptor-binding site around the newlyintroduced ligand, by which acceptor–ligand structure after mo-lecular docking can be refined by MD simulation (Alonso et al.,2006). The conformation changes of ACE/peptide complexesduring the MD simulation can be checked by analyzing thebackbone root mean square deviations (RMSD). Typically, thebackbone RMSD changes of ACE/VK and ACE/VKW complexwere analyzed. Both ACE/VK and ACE/VKW complexes wereequilibrated well after 5 ns of simulation, while the back-bone RMSD change of ACE/VKW complex was less than thatof ACE/VK complex (Supplementary Fig. S1), which suggestedthat the structure of ACE/VKW complex was more stable thanthat of ACE/VK complex.

Non-covalent interactions play an important role inbiomacromolecules, not only maintaining the three dimen-sional structure of large molecules but also are responsible forthe molecular recognition process (Cerny & Hobza, 2007). Theinteractions between peptides and ACE can be evaluated by in-teraction potential energies, which can be calculated accordingto the MD results (Lindahl et al., 2001). In Gromacs, the non-covalent interaction potential energy was evaluated by Coulombpotential energy and Lennard–Jones (LJ) potential energy. In thestudy, the total non-covalent interaction potential energybetween ACE and peptides (Etotal) was divided into three parts:Coulomb potential energy between ACE (exclude Zn) andpeptide (Ecoul) as electrostatic interaction, LJ potential energybetween ACE (exclude Zn) and peptide (ELJ) as van der Waalsforce, and Coulomb potential energy between Zn and peptide(EZn) as metal interaction. The interaction potential energiesbetween ACE and peptides were expressed as the average valueover the last 5 ns simulation (Table 3). Etotal of ACE/derivedpeptide complexes decreased compared with those of ACE/original peptide complexes, which showed that the stabilityof the ACE/peptide complexes can be strengthened by modi-fication with Trp. But each part of total potential energy changesbefore and after being modified with Trp depended on the

Table 2 – Docking scores from molecular docking of peptides into ACE.

Score Lisinopril VK VKW YA YAW KY KYW TAY TAW

Total Score 11.9 11.0 11.5 7.5 12.2 10.8 10.7 10.3 9.6*ΔTotal Score 0.5 4.7 −0.1 −0.7D_Score −158.4 −98.4 −158.8 −79.9 −130.6 −106.5 −178.0 −121.0 −129.4*ΔD_score 60.4 50.7 71.5 8.4G_Score −241.6 −155.0 −248.6 −175.1 −240.0 −203.0 −306.5 −202.2 −245.3*ΔG_Score 93.6 64.9 103.5 43.1Chemscore −37.8 −16.9 −27.9 −9.4 −24.2 −22.3 −26.9 −25.4 −29.1*ΔChemscore 11.0 14.8 4.6 3.7

* Δ the difference between absolute value of docking scores of derived peptides and original peptides; Total Score, D_score, G_score and Chemscoreare calculated by different algorithms in Sybyl.

635J o u rna l o f Func t i ona l F ood s 1 7 ( 2 0 1 5 ) 6 3 2 – 6 3 9

original peptide structure. For ACE/VK and ACE/VKW com-plexes, both Ecoul and ELJ decreased, but EZn increased after VKwas modified into VKW. For ACE/TAY and ACE/TAW complex,ELJ and EZn decreased, but Ecoul increased after TAY was modi-fied into TAW (Table 3).These results indicated that the stabilityof ACE/peptide complexes can be strengthened by modifica-tion with Trp at the C-terminus, but the mechanisms weredifferent depending on the original peptide structures. It hasbeen reported that Trp residues in different peptides show dif-ferent conformations and different interaction models in theACE/inhibitor complexes (Watermeyer, Kroger, O’Neill, Sewell,& Sturrock, 2010).

The interaction models of ACE/peptide complexes were ana-lyzed by Discovery Studio. To investigate the changes of ACE/peptide complexes before and after MD simulation, theinteraction model of ACE/VK and ACE/VKW complexes beforeand after MD was analyzed in detail (Figs 1 and 2). The amino

Table 3 – Average interaction potential energies betweenACE and peptides over the last 5 ns simulation.

Complex Ecoul

(kJ/mol)aELJ

(kJ/mol)bEZn

(kJ/mol)cEtotal

(kJ/mol)d

ACE/VK 5.15 −69.29 −387.45 −451.58ACE/VKW −165.74 −162.82 −184.52 −513.08ACE/YA −38.83 −77.65 −129.84 −246.31ACE/YAW −35.61 −183.78 −196.70 −416.08ACE/KY −113.05 −83.95 −255.25 −452.25ACE/KYW −73.58 −178.53 −218.62 −470.73ACE/TAY −104.90 −138.35 −6.04 −249.28ACE/TAW 39.54 −154.27 −372.59 −487.32

a Coulomb potential energy between ACE (exclude Zn) and inhibitors.b Lennard–Jones potential energy between ACE (exclude Zn) and

inhibitors.c Coulomb potential energy between Zn and inhibitors.d The total non-covalent interaction energy E = Ecoul + ELJ + EZn.

Fig. 1 – The interaction models of angiotensin-I converting enzyme (ACE)/VK complex before (a) and after (b) moleculardynamics simulation generated by Discovery Studio. Number before amino acid, the order number of amino acid residuesof ACE enzyme; ball, amino acid residues of ACE enzyme; gray line, the atoms of the peptide; orange line, pi interaction;green and blue dashed line, hydrogen bond; number on the dashed line, the bond length of the hydrogen bond. (Forinterpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

636 J o u rna l o f Func t i ona l F ood s 1 7 ( 2 0 1 5 ) 6 3 2 – 6 3 9

acid residues of ACE involved in the interaction with VK werealmost unchanged before and after MD simulation, but the in-teraction ways changed. Electrostatic interactions betweenALA356 and HIS383 with VK were changed into van der Waalsinteractions, while van der Waals interactions between HIS353,HIS387, GLU411 and ARG522 with VK were changed into elec-trostatic interactions after MD simulation (Fig. 1a and b). ForACE/VKW complex, the numbers of amino acid residue in-volving in van der Waals interaction were obviously reducedafter MD simulation (Fig. 2a and b). Moreover, the relative po-sition of Zn and VKW was greatly changed after the MDsimulation, which could lead to the change of metal interac-tion between VKW and ACE. MD simulation can get moreaccurate interaction models between peptides and ACE by aflexible way compared with molecular docking.

Interaction models of eight ACE/peptide complexes after MDsimulation were analyzed and the frequency of amino acid resi-dues of ACE involved in the interaction with all peptides (fourderived peptides and four original peptides) were counted,

respectively, to identify the important interaction sites (Fig. 3).The amino acid residues with frequency no less than four wereGLN281, ALA354, SER355, GLU376, VAL380, HIS383, GLU411,LYS511, HIS513, ARG522, TYR520 and TYR523, among whichGLN281, ALA354, SER355, LYS511, HIS513, TYR520 and TYR523play an important role for the ACE activity based on the crystalstructure of ACE/lisinopril complex (Natesh et al., 2003, 2004).By comparative analysis of the frequencies of amino acidresidue involved in the interaction with the original and derivedpeptides, the frequencies of GLN281, ALA354, GLU411, PHE457and LYS511 interacted with the derived peptide were higherthan with the original peptide, which showed that interac-tions between these amino acid residues and peptides playedan important role in activity improvement of derived peptides.

To further elucidate the action mechanisms of ACE/derivedpeptide complexes, the interaction model of ACE/VK and ACE/VKW complexes after MD simulation as well as ACE/lisinoprilcomplex (PDB: 1O86) was analyzed in detail (SupplementaryTable S2). The number of the amino acid residues involved in

Fig. 2 – The interaction models of angiotensin-I converting enzyme (ACE)/VKW complex before (a) and after (b) moleculardynamics simulation generated by Discovery Studio. Number before amino acid, the order number of amino acid residuesof ACE enzyme; ball, amino acid residues of ACE enzyme; gray line, the atoms of the peptide; orange line, pi interaction;green and blue dashed line, hydrogen bond; number on the dashed line, the bond length of the hydrogen bond. (Forinterpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

637J o u rna l o f Func t i ona l F ood s 1 7 ( 2 0 1 5 ) 6 3 2 – 6 3 9

the electrostatic interaction with VKW increased greatly com-pared to that withVK, including GLU162, GLN281, HIS383, LYS511,HIS513, TYR520 and ARG522, most of which were involved inthe interaction with lisinopril. The number of the amino acidresidues involved in the interaction did not change for van derWaals and reduced for hydrogen bond and pi–pi interaction.The electrostatic interactions between ACE and VKW played amore important role in keeping the stability of ACE/VKW com-plexes than the other interactions.The increase of ACE inhibitoryactivity of VKW compared with VK was mainly attributed tothe increase of electrostatic interactions between ACE and VKW.

4. Conclusion

The ACE inhibitory activities of four novel derived tripeptidesVKW, YAW, KYW and TAW by modification with Trp at theC-terminus were far higher than that of the corresponding origi-nal peptides, and did not show toxicity for the Chang cell line.The activity improvement of derived peptides was attributedto the increase of structure stability of ACE/derived peptide com-plexes by increasing binding affinity and interaction sites withimportant amino acid residues between ACE and derived pep-tides. Modification with Trp at the C-terminus was an effectiveway for designing novel ACE inhibitory peptides.The four novelderived peptides, especially VKW, had a potential use as func-tional ingredients and drug lead compounds.

Acknowledgements

This study was supported by the Fisheries Commercializa-tion Technology Development Program (Project No. 811001-3)

from the Ministry for Food, Agriculture, Forestry and Fisheries,Republic of Korea. We thank the National SupercomputingCenter (Shenzhen, China) for providing the computational re-sources and Sybyl (Surflex-Dock module).

Appendix: Supplementary material

Supplementary data to this article can be found online atdoi:10.1016/j.jff.2015.06.016.

R E F E R E N C E S

Acharya, K. R., Sturrock, E. D., Riordan, J. F., & Ehlers, M. R. (2003).ACE revisited: A new target for structure-based drug design.Nature Reviewers Drug Discovery, 2, 891–902.

Alonso, H., Bliznyuk, A. A., & Gready, J. E. (2006). Combiningdocking and molecular dynamic simulations in drug design.Medicinal Research Reviews, 26, 531–568.

Berendsen, H. J. C., Postma, J. P. M., van Gunsteren, W. F., &Hermans, J. (1981). Interaction models for water in relation toprotein hydration. Intermolecular forces. The JerusalemSymposia on Quantum Chemistry and Biochemistry, 14, 331–342.

Bussi, G., Donadio, D., & Parrinello, M. (2007). Canonical samplingthrough velocity rescaling. The Journal of Chemical Physics, 126,1–8.

Castellano, P., Aristoy, M., Sentandreu, M. A., & Vignolo, G. (2013).Peptides with angiotensin I converting enzyme (ACE)inhibitory activity generated from porcine skeletal muscleproteins by the action of meat-borne Lactobacillus. Journal ofProteomics, 89, 183–190.

Cerny, J., & Hobza, P. (2007). Non-covalent interactions inbiomacromolecules. Physical Chemistry Chemical Physics, 9,5291–5303.

Fig. 3 – The frequency of the amino acid residues of angiotensin-I converting enzyme (ACE) involved in the interaction withpeptides (the amino acid residues with frequency no less than 4 were listed when involved in interaction with all peptides).

638 J o u rna l o f Func t i ona l F ood s 1 7 ( 2 0 1 5 ) 6 3 2 – 6 3 9

Cushman, D. W., Cheung, H. S., Sabo, E. F., & Ondetti, M. A. (1977).Design of potent competitive inhibitors of angiotensin-converting enzyme. Carboxyalkanoyl and mercapto-alkanoylamino acids. Biochemistry, 16, 5484–5491.

De Leo, F., Panarese, S., Gallerani, R., & Ceci, L. R. (2009).Angiotensin converting enzyme (ACE) inhibitory peptides:Production and implementation of functional food. CurrentPharmaceutical Design, 15, 3622–3643.

Essmann, U., Perera, L., Berkowitz, M. L., Darden, T., Lee, H., &Pedersen, L. G. (1995). A smooth particle mesh Ewald method.The Journal of Chemical Physics, 103, 8577–8593.

Fields, G. B., Lauer-Fields, J. L., Liu, R., & Barany, G. (2002).Principles and practice of solid-phase peptide synthesis. In G.A. Grant (Ed.), Synthetic peptides: A user’s guide (pp. 93–219).New York: Oxford University Press, Inc.

Hess, B., Bekker, H., Berendsen, H. J. C., & Fraaije, J. G. E. M. (1997).LINCS: A linear constraint solver for molecular simulations.Journal of Computational Chemistry, 18, 1463–1472.

Holder, A., Birke, A., Eisele, T., Klaiber, I., Fischer, L., & Hinrichs, J.(2013). Selective isolation of angiotensin I-converting enzyme-inhibitory peptides from micellar casein and β-caseinhydrolysates via ultrafiltration. International Dairy Journal, 31,34–40.

Jorgensen, W. L., Chandrasekhar, J., Madura, J. D., Impey, R. W., &Klein, M. L. (1983). Comparison of simple potential functionsfor simulating liquid water. The Journal of Chemical Physics, 79,926–935.

Kellenberger, E., Rodrigo, J., Muller, P., & Rognan, D. (2004).Comparative evaluation of eight docking tools for dockingand virtual screening accuracy. Proteins, 57, 225–242.

Kulkarni, A., Han, Y., & Hopfinger, A. J. (2002). Predicting Caco-2cell permeation coefficients of organic molecules usingmembrane-interaction QSAR analysis. Journal of ChemicalInformation and Modelling, 42, 331–342.

Lindahl, E., Hess, B., & van der Spoel, D. (2001). Gromacs 3.0: Apackage for molecular simulation and trajectory analysis.Molecular Modeling Annual, 7, 306–317.

Majumder, K., Chakrabarti, S., Morton, J. S., Panahi, S., Kaufman,S., Davidge, S. T., & Wu, J. (2015). Egg-derived ACE-inhibitorypeptides IQW and LKP reduce blood pressure inspontaneously hypertensive rats. Journal of Functional Foods,13, 50–60.

Natesh, R., Schwager, S. L., Evans, H. R., Sturrock, E. D., &Acharya, K. R. (2004). Structural details on the binding ofantihypertensive drugs captopril and enalaprilat to humantesticular angiotensin I-converting enzyme. Biochemistry, 43,8718–8724.

Natesh, R., Schwager, S. L., Sturrock, E. D., & Acharya, K. R. (2003).Crystal structure of the human angiotensin-convertingenzyme–lisinopril complex. Nature, 421, 551–554.

Parrinello, M., & Rahman, A. (1981). Polymorphic transitions insingle crystals: A new molecular dynamics method. Journal ofApplied Physics, 52, 7182–7190.

Seki, E., Osajima, K., Matsufuji, H., Matsui, T., & Osajima, Y. (1995).Val-Tyr, and angiotensin-I converting enzyme inhibitor fromsardines that have resistance to gastrointestinal proteases.Journal of Agricultural Chemical Society of Japan, 69, 1013–1020.

Shamloo, M., Eck, P., & Beta, T. (2015). Angiotensin convertingenzyme inhibitory peptides derived from cereals. Journal ofHuman Nutrition & Food Science, 3, 1057–1066.

Silva, A. W. S. D., & Vranken, W. F. (2012). ACPYPE-Antechamberpython parser interface. BMC Research Notes, 5, 367–375.

Watermeyer, J. W., Kroger, W. L., O’Neill, H. G., Sewell, B. T., &Sturrock, E. D. (2010). Characterization of domain-selectiveinhibitor binding in angiotensin converting enzyme using anovel derivative of lisinopril. Biochemical Journal, 428, 67–74.

Wu, J., Aluko, R. E., & Muir, A. D. (2002). Improved method fordirect high-performance liquid chromatography assay ofangiotensin-converting enzyme-catalyzed reactions. Journal ofChromatography. A, 950, 125–130.

Wu, J., Aluko, R. E., & Nakai, S. (2006a). Structural requirements ofangiotensin I-converting enzyme inhibitory peptides:Quantitative structure-activity relationship study of di-andtripeptides. Journal of Agricultural and Food Chemistry, 54, 732–738.

Wu, J., Aluko, R. E., & Nakai, S. (2006b). Structural requirements ofangiotensin-I converting enzyme inhibitory peptides:Quantitative structure-activity relationship modeling ofpeptides containing 4–10 amino acid residues. QSAR &Combinational Science, 25, 873–880.

Wu, S., Feng, X., Lan, X., Xu, Y., & Liao, D. (2015). Purification andidentification of angiotensin-I converting enzyme (ACE)inhibitory peptide from lizard fish (Saurida elongata)hydrolysate. Journal of Functional Foods, 13, 295–299.

Xie, C. L., Kim, J. S., Ha, J. M., Choung, S. Y., & Choi, Y. J. (2014).Angiotensin-I converting enzyme inhibitor derived fromcross-linked oyster protein. Biomed Research International, 2014,379234. doi:10.1155/2014/379234.

Xing, L., Hodgkin, E., Liu, Q., & Sedlock, D. (2004). Evaluation andapplication of multiple scoring functions for a virtualscreening experiment. Journal of Computer-Aided MolecularDesign, 18, 333–344.

Yamada, A., Sakurai, T., Ochi, D., Mitsuyama, E., Yamauchi, K., &Abe, F. (2013). Novel angiotensin I converting enzymeinhibitory peptide derived from bovine casein. Food Chemistry,141, 3781–3789.

Yamada, Y., Matoba, N., Usui, H., Onishi, K., & Yoshikawa, M.(2002). Design of a highly potent anti-hypertensive peptidebased on ovokinin (2–7). Bioscience, Biotechnology, andBiochemistry, 66, 1213–1217.

Zhou, M., Du, K., Ji, P., & Feng, W. (2012). Molecular mechanism ofthe interactions between inhibitory tripeptides andangiotensin-converting enzyme. Biophysical Chemistry, 168,60–66.

639J o u rna l o f Func t i ona l F ood s 1 7 ( 2 0 1 5 ) 6 3 2 – 6 3 9