Intestinal transport of TRH analogs through PepT1: the role of in silico and in vitro modeling

Intestinal transport of TRH analogs throughPepT1: the role of in silico and in vitromodelingPravin Bagula†, Kailas S. Khomanea†, Siddharth S. Kesharwanib†,Preeti Pragyanb†, Prajwal P. Nandekarb†, Chhuttan Lal Meenac,Arvind K. Bansala, Rahul Jainc, Kulbhushan Tikood

and Abhay T. Sangamwarb*

The present study involves molecular docking, molecular dynamics (MD) simulation studies, and Caco-2 cell monolayerpermeability assay to investigate the effect of structural modifications on PepT1-mediated transport of thyrotropinreleasing hormone (TRH) analogs. Molecular docking of four TRH analogs was performed using a homology model ofhuman PepT1 followed by subsequent MD simulation studies. Caco-2 cell monolayer permeability studies of four TRHanalogswere performed at apical to basolateral and basolateral to apical directions. Inhibition experimentswere carriedout using Gly-Sar, a typical PepT1 substrate, to confirm the PepT1-mediated transport mechanism of TRH analogs. Pappof the four analogs follows the order: NP-1894<NP-2378<NP-1896<NP-1895. Higher absorptive transport wasobserved in the case of TRH analogs, indicating the possibility of a carrier-mediated transport mechanism. Further,the significant inhibition of the uptake of Gly-Sar by TRH analogs confirmed the PepT1-mediated transport mechanism.Glide docking scores of all the four analogues were in good agreement with their transport rates, suggesting the role ofsubstrate binding affinity in the PepT1-mediated transport of TRH analogs. MD simulation studies revealed that thepolar interactions with amino acid residues present in the active site are primarily responsible for substrate binding,and a downward trend was observed with the increase in bulkiness at the N-histidyl moiety of TRH analogs. Copyright© 2014 John Wiley & Sons, Ltd.

Keywords: Caco-2 cells; TRH analogs; molecular dynamics; molecular docking; PepT1-mediated transport; GastroPlus

INTRODUCTION

Oral drug delivery is the most preferred route of administration.Drugs can be primarily absorbed via an active and/or passivetransport mechanism. Structure property relationships havebeen extensively studied for passively transported drugs(Lipinski, 2004; Lipinski et al., 2012). The permeability of activelytransported drugs is mainly governed by their binding affinity to-ward the carrier or transporter proteins. Binding affinity involvesmolecular interactions between a drug and its transporter pro-tein. An understanding of these molecular interactions will helpus in establishing the structure property relationships. Recently,thyrotropin releasing hormone (TRH) has attracted the attentionof many researchers because of the neuropharmacological ac-tions in addition to its endocrine functions. TRH has been clini-cally proven effective in depression, epilepsy, spinocerebellardegeneration, amyotrophic lateral sclerosis, schizophrenia, andother central nervous system disorders (Monga et al., 2008a;Khomane et al., 2011b). However, its therapeutic use is limitedbecause of the poor permeability, short half-life, and endo-crine-related side effects. Recently, TRH analogs devoid of theendocrine related side effects and stable in the biological fluidincluding plasma are reported (Khomane et al., 2011b). However,the poor permeability of these analogs limits their deliverythrough oral route. Promising TRH analogs, azetirelin andMK-771, were withdrawn from clinical trials because of theirpoor permeability and bioavailability (Sasaki et al., 1994;Mahato et al., 2003).

* Correspondence to: A. T. Sangamwar, Department of Pharmacoinformatics,National Institute of Pharmaceutical Education and Research (NIPER), S.A.S.Nagar, Sector 67, Mohali, Punjab, India. PIN 160062.E-mail: [email protected]

† Authors have equal contribution.

a P. Bagul, K. S. Khomane, A. K. BansalDepartment of Pharmaceutics, National Institute of Pharmaceutical Educationand Research (NIPER), Sector-67, S.A.S. Nagar, Mohali, Punjab 160062, India

b S. S. Kesharwani, P. Pragyan, P. P. Nandekar, A. T. SangamwarDepartment of Pharmacoinformatics, National Institute of PharmaceuticalEducation and Research (NIPER), Sector-67, S.A.S. Nagar, Mohali, Punjab160062, India

c C. L. Meena, R. JainDepartment of Medicinal Chemistry, National Institute of PharmaceuticalEducation and Research (NIPER), Sector-67, S.A.S. Nagar, Mohali, Punjab160062, India

d K. TikooDepartment of Pharmacology and Toxicology, National Institute of Pharma-ceutical Education and Research (NIPER), Sector-67, S.A.S. Nagar, Mohali,Punjab 160062, India

Abbreviations: TRH, thyrotropin releasing hormone; MD, molecular dynamics;DMEM, Dulbecco’s modified Eagle’s medium; HBSS, Hank’s balanced salt solu-tions; FBS, fetal bovine serum; NEAAs, nonessential amino acids; trypsin-EDTA,trypsin–ethylenediaminetetraacetic acid; LY, Lucifer yellow; DMSO, dimethylsulfoxide; PBS, phosphate-buffered saline; MTT, 3-[4, 5-dimethylthiazol-2-yl]-2,5-diphenyl tetrazolium bromide; TEER, transepithelial electrical resistance.

Research Article

Received: 25 February 2014, Revised: 22 April 2014, Accepted: 22 April 2014, Published online in Wiley Online Library

(wileyonlinelibrary.com) DOI: 10.1002/jmr.2385

J. Mol. Recognit. 2014; 27: 609–617 Copyright © 2014 John Wiley & Sons, Ltd.

609

The literature survey provides the substantial evidence aboutthe TRH transport through the intestinal oligopeptide trans-porter, PepT1 (SLC15A1; Yokohama et al., 1984; Walter and Kissel,1994). In our previous study, we reported the transport andincreased binding of a TRH analog, NP-647, through PepT1 ascompared with the TRH (Khomane et al., 2012). Increased perme-ability of NP-647 over TRH was attributed to its increased bindingaffinity toward PepT1. Furthermore, the study of diverse analogsmay provide greater insights in the molecular binding with thePepT1.

Herein, we investigated the effect of structural modificationson the binding affinity of TRH analogs with PepT1. Previously, aseries of TRH analogs was synthesized and reported for theirantiepileptic activity (Jain et al., 2002; Kaur et al., 2005, 2006,2007; Monga et al., 2008b, 2011; Rajput et al., 2009a, 2009b,2011). The four potential analogs NP-1894, NP-1895, NP-1896,and NP-2378 (Figure 1) are investigated further for their PepT1-mediated transport mechanism. Molecular docking studies werecarried out to uncover the binding affinity of the TRH analogs withPepT1. Molecular interactions between TRH analogs and PepT1were investigated using molecular dynamics (MD) simulations.The PepT1-mediated transport mechanism was established usinga Caco-2 cell model. Transport studies of all the four analogs wereperformed at apical to basolateral (A→B) and basolateral to apical(B→A) directions. Competitive inhibition study was also carriedout using glycylsarcosine (Gly-Sar), a typical PepT1 substrate, toconfirm the PepT1-mediated transport mechanism.

MATERIALS AND METHODS

Molecular docking

The automated molecular docking of novel TRH analogues wasperformed in a PepT1 homology model using Glide softwareincluded in Schrodinger suite 9.0.02 (Friesner et al., 2004). Thehomology model of PepT1 was developed and validated asreported in our previous study (Khomane et al., 2012). The previ-ously validated molecular docking parameters (Khomane et al.,2012) were used for docking of TRH and its analogs, NP-1894,NP-2378, NP-1896, and NP-1895. The centroid of amino acidresidues His57, Phe293, Trp294, Leu296, and Phe297 was consid-ered as the center of receptor grid. A grid size of 30 Å was usedto accommodate the large tripeptide ligand in the active site ofPepT1 (Bolger et al., 1998; Meredith and Price, 2006). Moleculardocking was performed with standard precision using OptimizedPotentials for Liquid Simulations 2001 force field (Kaminski et al.,2001). A scaling factor of 0.8 with partial cutoff of 0.25 andCoulomb-vdW cutoff of 50 kcal/mol was used (Nandekar et al.,2013). The best-docked structure for each ligand was selectedon the basis of Glide docking score.

MD simulations

The MD simulations of the substrate-bound PepT1 complexeswas performed to calculate changes in binding free energy andper residue energy decomposition using AMBER 11 package(Case et al., 2005; Paesani et al., 2008). Binding free energy wascalculated using molecular mechanics–Poisson–Bolzmannsurface area (MM/PBSA) and MM/generalized Born SA MM/GBSAmethods. The MD simulations presented here model the behaviorof transporter–substrate complex under the conditions corre-sponding to those applied in the experimental studies. MD simula-tion studies were carried out as per the previously reportedprotocol (Khomane et al., 2012). Briefly, all hydrogens were addedusing the Leap program from the AMBER 11 package. The struc-tures were neutralized by adding Cl�counter ions in the solublecomplex. Each system was inserted in a rectangular TIP3P waterbox in which a protein atomwas at least 8 Å away from the nearestedge of the box. All simulations were run with SHAKE on hydrogenatoms, a 2-fs time step and Langevin dynamics for the temperaturecontrol. The system was energy minimized prior to the productionMD run in the following way. The protein was frozen and thesolvent molecules with counter ions were allowed to move during2000 steps of short minimization (including 1000 steps of steepestdescent followed by 1000 steps of conjugate minimization).After relaxation, the system was heated up to 300 K for 50 and50 ps of density equilibration with weak restraints (restraint_wt = 2.0) on the protein–ligand complex followed by 500 psof constant pressure equilibration under normal pressure andtemperature (NPT) conditions, that is, p = 1 atm, T = 300 K. Thesystem was equilibrated for 1 ns under periodic boundary con-ditions in the NPT ensemble at normal (300 K) temperaturesand a constant pressure of 1 atm using a 2-fs time integrationsteps, and trajectories were recorded after every 2-ps time step.The 9 Å of the cutoff was applied to treat the nonbonding inter-actions. The production MD simulation of 4 ns long was carriedout for all PepT1–ligand complexes. All MD trajectories were an-alyzed by the ptraj module of the AMBER 11 package and VisualMD (Humphrey et al., 1996). The change in binding free energyand per residue energy decomposition calculations wasperformed on the snapshots extracted from 4 ns of MD trajecto-ries using MM-PB/GBSA methods.

Chemicals

The TRH analogs NP-1894, NP-2378, NP-1896, and NP-1895were synthesized in the Department of Medicinal Chemistry,National Institute of Pharmaceutical Education and Research(NIPER), S.A.S Nagar, India, using solution phase peptidesynthesis. Gly-Sar, Dulbecco’s modified Eagle’s medium(DMEM), Hank’s balanced salt solutions (HBSS), fetal bovineserum (FBS)—heat inactivated, nonessential amino acids(NEAAs), trypsin–ethylenediaminetetraacetic acid (trypsin–EDTA) solution, penicillin–streptomycin–amphotericin solution,Lucifer yellow (LY), and dimethyl sulfoxide (DMSO) wereobtained from Sigma-Aldrich (St. Louis, MO, USA). 2-[4-(2-hydroxyethyl)-1-piperazinyl] ethanesulfonic acid, 2-(N-morpholino) ethanesulfonic acid, phosphate-buffered saline(PBS), 3-[4, 5-dimethylthiazol-2-yl]-2, 5-diphenyl tetrazolium-bromide (MTT), and sodium azide were acquired from HimediaLaboratories Pvt. Ltd (Mumbai, India). Verapamil was a gift fromNicholas Piramal India Ltd (Mumbai, India). Absolute ethanolwas procured from Hong Yang Chemical Co. Ltd (Jiangsu,

Figure 1. Chemical structure of TRH and general structure of TRH analogs.NP-1894 (R=benzyl), NP-2378 (R= isopropyl), NP-1896 (R=n-propyl), andNP-1895 (R=ethyl).

P. BAGUL ET AL.

wileyonlinelibrary.com/journal/jmr Copyright © 2014 John Wiley & Sons, Ltd. J. Mol. Recognit. 2014; 27: 609–617

610

China). Milli-Q grade water purified by a UV purification system(Millipore, Bedford, MA, USA) was used.

Cell culture

Caco-2 cells that originated fromAmerican Type Culture Collection,Manassas, VA, USA, at passage 37 were grown in DMEM with4500mg/l D-glucose, 110mg/l sodium pyruvate, and L-glutamineand supplemented with 15% of FBS, 1% penicillin–streptomycin–amphotericin solution, and 1% NEAA solution. Cells were cultured

in T 75-cm2 tissue culture flasks obtained from Cellstar®, GreinerBio-One (Frickenhausen, Germany). The cell cultures weremaintained at 37°C in a carbon dioxide (CO2) incubator, waterjacketed with HEPA Class 100 (Forma Series II, Thermo ElectronCorporation, USA) with the atmospheric air kept at 95% air and5% CO2 at 95% humidity. The cells became 80–85% confluent in4–7days after which they were harvested with trypsin-EDTA priorto seeding. The cells were grown on polycarbonate filters of 0.4-μ

pore size (Millicell® 24-well cell culture plate, Millipore, Billerica,MA, USA) at a seeding density of 7.5� 104 cells per well for21–22days to achieve a consistent monolayer. The growth mediawere changed, and the transepithelial electrical resistance (TEER)value was measured every alternate day. Cells from passagenumber 50–60 were used for the experiments.

MTT cytotoxicity assay

Cells were harvested and seeded in 96-well plates at a seedingdensity of 2� 104 cells per well. Solutions were prepared at aconcentration range of 25–300μM and incubated for 2 and 24h,respectively. Ten microliters of MTT solution (5mg/ml in PBS) wasthen added to each well and incubated for 4–5h (37°C, 5% CO2)to allow MTT to be metabolized. The media were dumped off,and formazan (metabolic product of MTT) was resuspended in100μl of DMSO and incubated for 1 h to enable thorough mixingof formazan into the solvent. Optical density was read at 560 nmusing ELISA Plate Reader (BioTek Instruments, Inc, USA), andbackground was subtracted at 670nm. The percent cell viabilitywas measured from Eqn (1).

Cell viability %ð Þ ¼ Signal� BackgroundBlank� Background

� 100 (1)

Stability study in HBSS

The solution of TRH analogs (1mM) in HBSS at pH 6.5 was kept ina shaker bath (37°C, 60 rpm) and analyzed using HPLC assaymethod at 0, 30, 60, 90, 120, and 180-min time points.

Permeability experiments

Permeability experiments were performed in a shaker incubatormaintained at 37°C and 60 rpm. The TEER value was measuredwith a Millipore Electrical Resistance System voltameter(Millipore, Billerica, MA, USA) in order to evaluate the monolayerintegrity. The TEER value was measured from Eqn (2).

TEER ¼ Rmonolayer � Rblank� �� A (2)

where Rmonolayer is the resistance of the cell monolayer alongwith the filter membrane, Rblank is the resistance of the filtermembrane, and A is the surface area of the membrane (0.7 cm2

in 24-well plates).

The transport studies of TRH analogs and Gly-Sar (1mM) wereconducted at apical to basolateral (A→ B) and basolateral toapical (B→A) directions using a gradient method, that is, pH6.5 at apical and pH 7.4 at basolateral side. For A→ B studies,400μl of the drug solution in HBSS was added to the apical side(maximum chamber volume 800μl), and 800μl of the blanktransport buffer was added to the basolateral side (maximumchamber volume 1100μl). A sample volume of 600μl waswithdrawn from the basolateral side at 15, 30, 45, 60, 90, and120min. The volume withdrawn was replaced with blank trans-port buffer each time. In B→A studies, the initial solution wasadded to the basolateral side, and the concentration in the apicalside was measured (Sun and Pang, 2008; Wahlang et al., 2011;Khomane et al., 2012; Bagul et al., 2014). The apparent perme-ability coefficients, Papp (cm/s), for both A→ B and B→A studieswere calculated from Eqn (3).

Papp ¼ dQdt

=C0 A (3)

where dQ/dt is the cumulative transport rate (μM/min) definedas the slope obtained by the linear regression of cumulativetransported amount as a function of time (min), A is the surfacearea of the filters or inserts (0.7 cm2 in 24 wells), and C0 is theinitial concentration of the compounds on the donor side (μM).The efflux ratio (ER) was calculated from the following equation:

ER ¼ Papp ABð Þ=Papp BAð Þ (4)

Inhibition studies

In inhibition experiment, 1mM of each TRH analog (inhibitor)was dissolved in transport medium containing 200μM ofGly-Sar and applied to the apical side of the monolayer. Theinhibitory effect of each TRH analog on Gly-Sar transport acrossa Caco-2 monolayer was calculated as % inhibition using Eqn (5).

% inhibition ¼ Pwithout inhibitor � Pwith inhibitor

Pwithout inhibitor� 100 (5)

where P indicates permeability value of Gly-Sar in cm/s.

Monolayer integrity test

At the end of each experiment, the monolayer integrity test wasperformed to analyze the concentration of LY in the apical andbasolateral compartments. An initial stock solution of LY(50mM) was prepared and diluted to a 100-μM working solution.Working solution (400μl) of LY was added to the apical side ofthe Caco-2 cell monolayer (in the wells in which drug transportstudy was performed), and 800μl of HBSS buffer (pH 7.4) wasadded to the basolateral side. The plate was then kept in ashaker incubator at 37°C and 60 rpm. After 120min, 700 and300μl of the samples were withdrawn from the basolateral sideand apical side, respectively. The samples were analyzed by fluo-rescence spectroscopy at an excitation wavelength of 485 nmand emission wavelength of 535 nm using a spectrofluo-rophotometer (RF-5301-PC, Shimadzu, Japan).

HPLC analysis

HPLC analysis of TRH analogs was carried out using a reversedphase LiChrospher C8 analytical column (4.6� 250mm, 5-μmparti-cles; Merck, KgaA, Darmstadt, Germany), maintained at 35°C. The

ROLE OF STRUCTURAL MODIFICATIONS ON PEPT1-MEDIATED TRANSPORT

J. Mol. Recognit. 2014; 27: 609–617 Copyright © 2014 John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/jmr

611

column was preceded by a 3� 4mm C8 guard column(LichroCART, Merck). The mobile phase consisted of a mixture oforganic phase (acetonitrile) and aqueous phase (1% TFA buffer,pH 2.2) in the ratio of 72:28, respectively. The flow rate and detec-tion wavelength (λmax) were 1ml/min and 272nm, respectively.

Partition coefficient determination

Partition coefficient (log P) was determined as per OECD guide-lines for the testing of chemicals (Klein et al., 1988; Khomaneet al., 2011a). pKa values of the four analogs were predictedusing ACD/I-lab, v11.0 (ACD/I-Lab Service, Canada). Based onpredicted pKa values (6.29, 6.65, 6.59, and 6.58 for NP-1894,NP-2378, NP-1896, and NP-1895, respectively), Tris buffer of pH9.0 was selected as an aqueous phase. All TRH analogs were inunionized state at pH 9.0. High purity analytical grade n-octanolwas used as an organic phase. Organic phase, n-octanol wasshaken with sufficient quantity of the aqueous phase for 24 hon a mechanical shaker at 100 rpm to saturate the solvent witheach other and was allowed to stand till the different phaseswere separated. Accurately weighed TRH analog was then addedto the aqueous phase presaturated with n-octanol and wasfollowed by sonication to achieve a solution of 1 μg/ml.An experiment was carried out in duplicate using three differentvolume ratios of 1:1 (1.5ml each phase), 1:2 (1ml n-octanol and2ml buffer), and 2:1 (2ml n-octanol and 1ml buffer). A volumet-ric flask (3ml) was nearly filled by the entire volume of the twophases to prevent the loss of the material as a result of volatiliza-tion. Fifteen piston strokes were applied to the partition mediumusing a 10-ml glass syringe to achieve uniform dispersion of twophases, and then, it was kept in shaker water bath maintainedat 37±1°C for 3h at 100 rpm. Separation of phases was achievedby centrifugation at 3000 rpm for 10min. An aliquot of eachof the phases was taken for analysis. HPLC analysis was performedin triplicate using the previously mentioned validated HPLCmethod. Equation (6) was used to calculate the log P values.

LogP ¼ LogCo

Caq

� �(6)

where Co and Caq are the concentration of TRH analogs inn-octanol and aqueous phases, respectively. Caq was takento minimize the risk of the presence of traces of n-octanolwhile sampling the aqueous phase.

RESULTS

Molecular docking

Molecular docking studies revealed that amino acid residuesThr58, Trp294, Gln300, Thr625, Val628, Gly629, Gly638, Ala639,and Glu648 were involved in the hydrogen bonding interactionswith polar groups of NP-1894, NP-2378, NP-1896, and NP-1895ligands. These H-bonding interactions were reported as specificinteractions for the PepT1 substrate (Meredith and Price, 2006).The surrounding amino acid residues interacting with thedocked TRH analogs were Trp47, Leu51, Ser52, Thr53, Ile55,His57, Thr58, Ala61, Trp294, Asp298, Gln300 Thr625, Val626,Ala627, Val628, Gly629, Asn630, Ile631, Ile632, Val633, Ile635,Val636, Ala637, Gly638, Ala639, Gly640, Lys644, Glu648, Tyr649,Leu651, Phe652, and Leu655. The amino acid residue Phe652was involved in pi-stacking interaction with a pyrazine moietyof tripeptide substrate. On the other hand, pi-cationic interaction

was observed between the histidyl group of the substrate andLys644. These observations are consistent with the resultsreported by Meredith et al. and Khomane et al. in earlier studies(Meredith and Price, 2006; Khomane et al., 2012). The Glidedocking score of the four analogs follows the order NP-1894(�8.03)<NP-2378 (�8.06)<NP-1896 (�8.45)<NP-1895 (�8.65).The docking score indicates the binding affinity of these analogstoward PepT1. The observed docking score follows the sameorder as the experimental Papp values determined using theCaco-2 cell monolayer as summarized in Table 1.

MD simulations

PepT1–substrate complexes obtained from molecular dockingwere used as starting structure for MD simulation run. The 4 nslong MD simulation of all PepT1–substrate complexes showedthat the geometry of the PepT1–substrate complex was wellmaintained. The root-mean-square deviation (RMSD) plots forPepT1–ligand complexes and ligands were shown in Figures 2Aand 2B, respectively, and were found to converge by the endof 4-ns MD simulations. The RMSD plots of PepT1–ligand com-plexes showed an initial increase and then plateau, implying thatthe system has folded up to a state more stable than the startinglinear structure.The binding free energy (ΔG binding) calculations were

performed using the MM-PB/GBSA method. The average per-residuewise energy decomposition was calculated for the last2-ns MD simulation run. Glide docking score, MM-GBSA bindingfree energy, and E-GB/PB polar contribution were used to com-pare the PepT1–substrate binding as shown in Table 1. The totalMM-GBSA binding free energy for NP-1894, NP-2378, NP-1896,and NP-1895 were found to be �24.76, �34.15, �39.27, and�41.79 kcal/mol, respectively. This followed the same order asthat of the docking score thereby supporting the moleculardocking results.The polar interaction energy contribution (EGB) to the total ΔG

binding energy for four analogs was estimated as 39.48, 46.42,49.87, and 54.05 kcal/mol, respectively. These values indicatedthat the polar interactions were responsible for tight binding ofthe substrates to PepT1. The estimated contribution of polarinteractions to total binding free energy showed up with approx-imately a 15 kcal/mol increase in polar energy for NP-1895 ascompared with NP-1894. The polar energy contribution wasmainly because of polar amino acid residues present in the ac-tive site of PepT1. The H-bonding interactions of TRH analogswith these polar amino acid residues were primarily responsiblefor substrate binding and hence their PepT1-mediated transport.The amino acid residues involved in hydrogen bonding inter-

actions with these substrates are shown in Figure 3. As shown inFigure 3A, the two inconsistent H-bonding interactions wereobserved between PepT1 and NP-1894 at initial frames of MDsimulation run. Thus, it pointed toward less or unfavorable bind-ing of NP-1894 that lead to its lowest binding affinity and trans-port among the four TRH analogs. This was further supported byhigher RMSD of NP-1894 in the 4-ns MD simulation run as shownin Figure 2B, revealing higher flexibility of the ligand (NP-1894)and less binding stability at the active site of PepT1, while inFigure 3B, NP-2378 showed two H-bonding interactionswith amino acid residues Thr58, Val628, and Glu648. Out ofthe three H-bonding interactions, the H-bonding interactionwith polar amino acid residue Glu648 was found to be consis-tent throughout the MD simulation run. Similarly, in the case

P. BAGUL ET AL.


612

of analog NP-1896, the H-bonding interactions with polar resi-dues Thr58 and Glu648 were found to be consistent as shownin Figure 3C.The moderate range of MM-GBSA binding energy of NP-2378

(�34.15 kcal/mol) and NP-1896 (�39.27 kcal/mol) resultedbecause of the presence of these observed H-bonding interac-tions. The stable interaction pattern was also supported by lowerRMSD for these ligand confirmations in their docking posethroughout the MD simulation run as shown in Figure 2B. Thissuggested the higher and stable binding of NP-2378 andNP-1896 with PepT1 that contributed in their higher Caco-2transport rate as compared with NP-1894.The molecular docking and MD simulation results of NP-1895

suggested that the stable and consistent H-bonding interactionswere observed with amino acid residues Gly629, Val628, andGlu648 as shown in Figure 3D. The RMSD profile for ligandconfirmation in its bound state was also found to be stable in

the MD simulation run. This offered NP-1895 the highest bindingfree energy and hence greater affinity as compared with otheranalogs. On the whole, the previous observations support thefact that PepT1-mediated transport of TRH analogs could begoverned by the substrate binding affinity that was, in turn,influenced by substrate interactions with polar amino acid resi-dues present in the active site.

The involvement of crucial amino acid residues in the bindingof substrates with PepT1 was further analyzed using a per-residue decomposition energy profile as shown in Figure 4(Ahmad et al., 2009). The amino acid residues Trp47, Leu51,Ser52, Thr53, His57, Thr58, Trp294, Thr625, Ala627, Val628,Gly629, Asn630, Gly638, Ala639, Glu648, Tyr649, Leu651, andPhe652 were found to have significant contribution towardPepT1–substrate binding. It was observed that the polar interac-tions of amino acid residues His57, Ala627, Gly629, Glu648,Tyr649, and Phe652 with NP-1895 were stronger than the other

Table 1. Various experimental and in silico parameters for TRH analogs

Compound Papp Caco-2

(10�6 cm/s)aEffluxratio

Log P SimPeff human

(10�6 cm/s)bGlide dockingscore (kcal/mol)

MM-GBSA ΔGc

binding(kcal/mol)

EGB(polar contribution;

kcal/mol)

NP-1894 11.99 (0.20) 1.30 �0.14 (0.11) 249.50 �8.03 �24.76 39.78NP-2378 21.97 (1.07) 1.38 �0.72 (0.13) 345.88 �8.06 �34.15 46.42NP-1896 28.72 (1.67) 1.90 �0.45 (0.10) 399.83 �8.45 �39.27 49.87NP-1895 30.23 (0.50) 1.96 �1.01 (0.11) 411.07 �8.65 �41.79 54.05

Standard deviations are given in parentheses.MM-GBSA delta G (ΔG) binding, final estimated binding free energy (kcal/mol) calculated from MD simulation; EGB = theelectrostatic contribution to the solvation free energy (kcal/mol) calculated by GB.aApparent permeability coefficients in the Caco-2 cell monolayer;bSimulated effective permeability in humans;cFree energy

Figure 2. Graph showing RMSD changes throughout the 1-ns equilibration and 4-ns production simulation run for (A) PepT1–ligand complexes and(B) ligands in its bound confirmations.



613

analogs. Similarly, in the case of NP-1894, NP-2378, and NP-1896,the per-residuewise energy decomposition contribution of polaramino acid residues, namely, Ser52, Thr53, Thr58, Trp294,Thr625, Asn630, Gly638, and Phe652 were significant ascompared with other amino acid residues. It clearly indicatesthat polar interactions of the mentioned amino acid residuesare crucial for the substrate recognition by PepT1. The PepT1–substrate binding energy was influenced by these polar aminoacid residues that ultimately influenced the substrate transportrate. This was further validated using in vitro permeability exper-iments across the Caco-2 cell monolayer.

Permeability experiments

TRH analogs were found to be stable in HBSS buffer during thetime period required for the permeability studies that is at leastfor 2 h. The RSD of the peak area of all the TRH analogs wasfound to be less than 2%. Viability of cells was directly measuredusing the MTT test to evaluate the cytotoxicity of TRH analogs onCaco-2 cells. TRH analogs when incubated with Caco-2 cells overa concentration range of 0.1 to 3mM showed no reduction in

cell viability, thereby indicating their nontoxic nature. Transportstudies of TRH analogs and Gly-Sar were performed using theCaco-2 cell model, and the permeability values are shown inTable 1. TRH analogs showed higher Papp values indicating theirimproved permeability over TRH (9.17� 10�6 cm/s) (Khomaneet al., 2012). NP-1895 had shown the highest Papp(30.23 ± 0.50� 10�6 cm/s) among the four TRH analogs. Thepermeability of the four analogues displayed the following order:NP-1894<NP-2378<NP-1896<NP-1895. TRH analogs showedsignificantly higher absorptive transportation than secretive,thus indicating a possibility of carrier-mediated transport. HigherERs (Table 1) also supported the involvement of an active com-ponent in the transport of TRH analogs. The involvement ofPepT1 in the transport of these TRH analogs was speculatedand further investigated using an inhibition study. The inhibitoryeffect of TRH analogs on the transport of Gly-Sar, a typical PepT1substrate, was investigated (Khomane et al., 2012).The Papp values of Gly-Sar at A→B and B→A directions were

found to be 58.11±1.19� 10�6 cm/s and 31.60±0.70� 10�6 cm/s,respectively. As shown in Figure 5, TRH analogs significantlyinhibited the uptake of Gly-Sar. The uptake of Gly-Sar (200μM)

Figure 3. Substrate binding pose and H-bonding interaction pattern between PepT1 and (A) NP-1894, (B) NP-2378, (C) NP-1896, and (D) NP-1895. Hbonds are shown in yellow color. Ligands are represented in pink colored stick view, and active site amino acid residues are represented in greencolored line view. Images were taken from PyMOL v1.5. Hydrogen bonding definition: distance between donor (D) and acceptor (A) atom< 3.2 Å; anglebetween D-H-A> 90°.

P. BAGUL ET AL.


614

was decreased by almost 40, 42, 44, and 46% in the presence of1mM of NP-1894, NP-2378, NP-1896, and NP-1895, respectively.Additionally, the effective permeability of these analogs in humanswas predicted by computer simulations performed using the

permeability model of GastroPlus™version 8.0.002 (Simulations Plus,

Inc., Lancaster, USA). The Papp values for TRH analogs obtained fromCaco-2 experiments were used as starting values (input) to simulatethe effective permeability in humans as expressed by SimPeffhuman. It was observed that the SimPeff human values for TRHanalogs followed the similar trend as that of Papp as shown in Table 1.

Partition coefficient determination

The log P values obtained for TRH analogs are reported inTable 1. All TRH analogs showed log P values less than zero(negative), indicating their hydrophilic nature. Out of the fouranalogues, NP-1894 showed the highest log P value that wasattributed to its benzyl substitution at the N-histidyl group. LogP values of the four analogs exhibited the following order: NP-1894 (benzyl)>NP-1896 (n-propyl)>NP-2378 (isopropyl)>NP-1895 (ethyl).

DISCUSSION

In the case of carrier-mediated transport, binding affinity to-ward the active site is a decisive factor. Just like the hydropho-bicity of a compound, the binding affinity toward PepT1 can

also be used as a screening parameter for the drugs absorbedthrough PepT1-mediated transport. The effect of structuralmodifications on passive transport of molecules is extensivelystudied (Ribadeneira et al., 1996). However, how these chemicalmodifications influence the PepT1-meditated transport alsodeserves an attention. The molecular docking and MD simula-tion analysis of atomic level interactions between PepT1 andTRH analogs (substrate) indicated that the substrate bindingwas influenced by the presence of polar amino acid residuesin the active site of PepT1. A good correlation exists betweensubstrate transport (indicated by Papp values) and binding affin-ity toward PepT1 (as indicated by the Glide docking score andbinding free energy). As the binding affinity of PepT1–substrateincreases, the extent of PepT1-mediated transport alsoincreased (Table 1).

Molecular interactions of TRH analogs NP-1894, NP-2378,NP-1896, and NP-1895 with PepT1 were found to be altered bysubstitutions at the N-histidyl. As size of N-substituted groupincreases, substrate interactions with polar amino acid residuesof the active site were found to decrease, resulting in reducedbinding affinity toward PepT1. In the case of NP-1895, Glu648forms an H bond with the amino group of the pyrazinamidemoiety, and this H bond is consistent throughout the MD simu-lation run as shown in Figure 3D. NP-1896 has n-propyl substitu-tion at the N-histidyl moiety; here, Thr58 forms an H bond withthe carbonyl group of pyrazinamide. The benzyl substitution atN-histidyl in NP-1894 leads to a change in the hydrogen bondingpattern that now shifts to Thr58, forming an H bond with thecarbonyl group of pyrazinamide of NP-1894, while Gly638 formsan H bond with the carbonyl group of prolinamide. The isopropylgroup at the N-histidyl of NP-2378 leads to a new H-bondingpattern, where Thr58 and Glu648 form an H bond with thecarbonyl group and amino group of pyrazinamide, respectively.Moreover, Val628 forms an H bond with the carbonyl group ofthe prolinamide moiety of NP-2378. Isopropyl substitutionreduces the intensity of H bonding as compared with that ofThr58 in NP-1896. The higher EGB value for NP-1896 as com-pared with NP-2378 supports its higher binding affinity. Theobservations from four interaction energy calculating modelsand per residue decomposition energy contribution suggeststhat an increase in bulkiness at the N-histidyl group of TRH ana-logs leads to a decrease in polar interactions with the previouslymentioned amino acid residues. These results into a decrease inPepT1-mediated transport. The previous observations are inagreement with the results obtained from in vitro Caco-2 cellpermeability studies. Thus, the present study suggests that the

Figure 4. Residuewise decomposition energy for active site amino acid residues for the mentioned substrates in the PepT1 homology model. Aminoacid residues are represented in single letter code.

Figure 5. Cumulative transport of Gly-Sar in the absence and in thepresence of TRH analogs (1mM) with respect to time.



615

structural changes favoring polar interactions may enhance thebinding affinity of the TRH analogs toward PepT1 and subse-quently PepT1-mediated transport of drugs.

The binding affinity of TRH analogs toward PepT1 was investi-gated successfully. It was found that all the four TRH analogshave much higher permeation across the Caco-2 monolayer ascompared with TRH. The higher absorptive transport than secre-tive transport clearly indicates the carrier-mediated transportof TRH analogs. TRH analogs significantly inhibited the uptakeof Gly-Sar, a typical substrate of PepT1. This confirmed therole of PepT1 in the transport mechanism of TRH analogs. Inmost of the carrier-mediated absorption of drugs, both the com-ponents, carrier mediated (saturable) and passive (nonsaturable)component, are involved. ER can also be used to study the rela-tive contribution of the active and passive component. A higherER indicates greater contribution of the carrier-mediated satura-ble component. In the present study, TRH analogs showed signif-icantly higher ERs. These ERs correlated well with the bindingaffinities of the four analogs represented by their Glide dockingscores and binding free energy. The ER increases with anincrease in drug binding affinity toward PepT1. All four TRHanalogs showed higher log P values as compared with TRH(log P =�2.46 (Bundgaard and Moss, 1990)); however, theirhigher permeability across the Caco-2 cell monolayer could notbe attributed to their increased hydrophobicity. It is because therewas no correlation observed between log P and Papp values of allthe four analogs as shown in Table 1. On other hand, this was inagreement with the carrier-mediated transport of the drug wherebinding affinity rather than hydrophobicity governs the transport.

CONCLUSIONS

The present work investigates the effect of structural modifica-tion on the transport mechanism of TRH analogs. The reported

analogs were found to be substrates for PepT1 that exhibitedhigher ERs and significantly inhibited the Gly-Sar transportacross the Caco-2 cell line. Molecular docking and MD simulationstudies revealed that the PepT1-mediated transport of theseanalogs was governed by their binding affinity toward the PepT1active site. Glide docking scores and binding free energiescorrelated well with the experimental Papp values of these TRHanalogs. The H-bonding interactions of these TRH analogs withthe polar amino acid residues present in the PepT1 active sitewere primarily responsible for the substrate recognition andbinding affinity toward PepT1. The per-residuewise energydecomposition analysis identified the important amino acidresidues responsible for these interactions. It was observed thatthese polar interactions were decreased with an increase inbulkiness at N-histidyl of TRH analogs. From the present study,we concluded that the binding affinity and polar interactionparameters could suitably explain the PepT1-mediated intestinalabsorption of TRH analogs. The previously employed parameterscan be used for primary screening of compounds at drug devel-opment phase. In the future, the extension of these studies formore diverse compounds could provide better understandingof PepT1–substrate interactions and more screening parametersto pharmacokinetic scientists. These could be useful for design-ing better drugs.

CONFLICT OF INTEREST

The authors have no conflict of interest.

Acknowledgement

The authors acknowledge financial support from NIPER S.A.S.Nagar and Department of Biotechnology, New Delhi.

REFERENCES

Ahmad R, Brandsdal BO, Michaud-Soret I, Willassen NP. 2009. Ferricuptake regulator protein: Binding free energy calculations and per-residue free energy decomposition. Proteins 75(2): 373–386.

Bagul P, Khomane KS, Bansal AK. 2014. Investigating Permeability RelatedHurdles in Oral Delivery of 11-Keto-β-Boswellic Acid. Int. J. Pharm.464(1–2): 104–110.

Bolger MB, Haworth IS, Yeung AK, Ann D, von Grafenstein H, Hamm-AlvarezS, Okamoto CT, Kim KJ, Basu SK, Wu S. 1998. Structure, function, andmolecular modeling approaches to the study of the intestinal dipep-tide transporter PepT1. J. Pharm. Sci. 87(11): 1286–1291.

Bundgaard H, Moss J. 1990. Prodrugs of peptides. 6. bioreversible deriv-atives of thyrotropin-releasing hormone (TRH) with increased lipo-philicity and resistance to cleavage by the TRH-specific serumenzyme. Pharm. Res. 7(9): 885–892.

Case DA, Cheatham TE, III, Darden T, Gohlke H, Luo R, Merz KM, Jr,Onufriev A, Simmerling C, Wang B, Woods RJ. 2005. The Amber bio-molecular simulation programs. J. Comput. Chem. 26(16): 1668–1688.

Friesner RA, Banks JL, Murphy RB, Halgren TA, Klicic JJ, Daniel T, RepaskyMP, Knoll EH, Shelley M, Perry JK. 2004. Glide: a new approach forrapid, accurate docking and scoring. 1. Method and assessment ofdocking accuracy. J. Med. Chem. 47(7): 1739–1749.

Humphrey W, Dalke A, Schulten K. 1996. VMD: visual molecular dynamics.J. Mol. Graph. 14(1): 33–38.

Jain R, Singh J, Perlman JH, Gershengorn MC. 2002. Synthesis and biologyof new thyrotropin-releasing hormone (TRH) analogues. Bioorg. Med.Chem. 10(1): 189–194.

Kaminski GA, Friesner RA, Tirado-Rives J, Jorgensen WL. 2001. Evaluationand reparametrization of the OPLS-AA force field for proteins viacomparison with accurate quantum chemical calculations on pep-tides. J. Phys. Chem. B 105(28): 6474–6487.

Kaur N, Lu X, Gershengorn MC, Jain R. 2005. Thyrotropin-releasinghormone (TRH) analogues that exhibit selectivity to TRH receptorsubtype 2. J. Med. Chem. 48(19): 6162–6165.

Kaur N, Monga V, Josan JS, Lu X, Gershengorn MC, Jain R. 2006. Synthesis,receptor binding, and activation studies of N (1)-alkyl-l-histidinecontaining thyrotropin-releasing hormone (TRH) analogues. Bioorg.Med. Chem. 14(17): 5981–5988.

Kaur N, Monga V, Lu X, Gershengorn MC, Jain R. 2007. Modifica-tions of the pyroglutamic acid and histidine residues inthyrotropin-releasing hormone (TRH) yield analogs with selec-tivity for TRH receptor type 2 over type 1. Bioorg. Med. Chem.15(1): 433–443.

Khomane K, Kumar L, Meena C, Jain R, Bansal A. 2011a. NP-647, a novelTRH analogue: Investigating physicochemical parameters critical forits oral and parenteral delivery. Int. J. Pharm. 406: 21–30.

Khomane KS, Meena CL, Jain R, Bansal AK. 2011b. Novel thyrotropin-releasing hormone analogs: a patent review. Expert Opin. Ther. Pat.21(11): 1673–1691.

Khomane KS, Nandekar PP, Wahlang B, Bagul P, Shaikh N, Pawar YB,Meena CL, Sangamwar AT, Jain R, Tikoo K, Bansal AK. 2012. Mecha-nistic Insights into PEPT1-Mediated Transport of a NovelAntiepileptic, NP-647. Mol. Pharm. 9(9): 2458–2468.

Klein W, Kördel W, Weiss M, Poremski H. 1988. Updating of the OECDtest guideline 107 “partition coefficient n-octanol/water”: OECDlaboratory intercomparison test on the HPLC method. Chemosphere17(2): 361–386.

Lipinski CA. 2004. Lead-and drug-like compounds: the rule-of-five revolu-tion. Drug Discov. Today Tech. 1(4): 337–341.

Lipinski CA, Lombardo F, Dominy BW, Feeney PJ. 2012. Experimental andcomputational approaches to estimate solubility and permeability in

P. BAGUL ET AL.


616

drug discovery and development settings. Adv. Drug Deliv. Rev. 64:4–17.

Mahato RI, Narang AS, Thoma L, Miller DD. 2003. Emerging trends in oraldelivery of peptide and protein drugs. Crit. Rev. Ther. Drug CarrierSyst. 20(2–3): 153–214.

Meredith D, Price RA. 2006. Molecular modeling of PepT1 - towards astructure. J. Membr. Biol. 213(2): 79–88.

Monga V, Meena CL, Kaur N, Jain R. 2008a. Chemistry and biology ofthyrotropin-releasing hormone (TRH) and its analogs. Curr. Med.Chem. 15(26): 2718–2733.

Monga V, Meena CL, Kaur N, Kumar S, Pawar C, Sharma SS, Jain R. 2008b.Facile Synthesis of Na-Boc-1, 2-Dialkyl-L-histidines: Utility inthe Synthesis of Thyrotropin-Releasing Hormone (TRH) Ana-logues and Evaluation of the CNS Activity. J. Heterocycl. Chem.45: 1603–1608.

Monga V, Meena CL, Rajput S, Pawar C, Sharma SS, Lu X, GershengornMC, Jain R. 2011. Synthesis, receptor binding, and CNS pharmacolog-ical studies of new Thyrotropin Releasing Hormone (TRH) analogues.Chem. Med. Chem. 6(3): 531–543.

Nandekar PP, Tumbi KM, Bansal N, Rathod VP, Labhsetwar LB, Soumya N,Singh S, Sangamwar AT. 2013. Chem-bioinformatics and in vitroapproaches for candidate optimization: a case study of NSC745689as a promising antitumor agent. Med. Chem. Res. 22(8): 3728–3742.

Paesani F, Vanicek J, Wu X, Brozell SR, Steinbrecher T, Gohlke H, Yang L,Tan C, Mongan J, Hornak V. 2008. AMBER 10. University of California:San Francisco.

Rajput SK, Krishnamoorthy S, Pawar C, Kaur N, Monga V, Meena CL, Jain R,Sharma SS. 2009a. Antiepileptic potential and behavioral profile ofl-pGlu-(2-propyl)–l-His–l-ProNH2, a newer thyrotropin-releasinghormone analog. Epilepsy Behav. 14(1): 48–53.

Rajput SK, Singh JN, Ingole S, Jain G, Kaur N, Monga V, Meena CL, Jain R,Sharma SS. 2009b. Neuropharmacological profile of l-pGlu-(1-ben-zyl)-l-His-l-ProNH2, a newer thyrotropin-releasing hormone analog:Effects on seizure models, sodium current, cerebral blood flow andbehavioral parameters. Epilepsy Res. 87(2): 223–233.

Rajput SK, Siddiqui MA, Kumar V, Meena CL, Pant AB, Jain R, Sharma SS.2011. Protective effects of L-pGlu-(2-propyl)-L-His-L-ProNH2, a newerThyrotropin Releasing Hormone analog in in vitro and in vivo modelsof cerebral ischemia. Peptides 32(6): 1225–1231.

Ribadeneira MD, Aungst BJ, Eyermann CJ, Huang SM. 1996. Effects ofstructural modifications on the intestinal permeability of angiotensinII receptor antagonists and the correlation of in vitro, in situ, andin vivo absorption. Pharm. Res. 13(2): 227–233.

Sasaki I, Fujita T, Murakami M, Yamamoto A, Nakamura E, Imasaki H,Muranishi S. 1994. Intestinal absorption of azetirelin, a newthyrotropin-releasing hormone (TRH) analogue. I: possible factors forthe low oral bioavailability in rats. Biol. Pharm. Bull. 17(9): 1256–1261.

Sun H, Pang KS. 2008. Permeability, transport, and metabolism ofsolutes in Caco-2 cell monolayers: a theoretical study. Drug Metab.Dispos. 36(1): 102–123.

Wahlang B, Pawar YB, Bansal AK. 2011. Identification of permeability-related hurdles in oral delivery of curcumin using the Caco-2 cellmodel. Eur. J. Pharm. Biopharm. 77: 275–282.

Walter E, Kissel T. 1994. Transepithelial transport and metabolism ofthyrotropin-releasing hormone (TRH) in monolayers of a humanintestinal cell line (Caco-2): evidence for an active transport compo-nent? Pharm. Res. 11(11): 1575–1580.

Yokohama S, Yoshioka T, Yamashita K, Kitamori N. 1984. Intestinalabsorption mechanisms of thyrotropin-releasing hormone.J. Pharmacobiodyn. 7: 445–450.



617

Documents

Intestinal transport of TRH analogs through PepT1: the role of in silico and in vitro modeling