Interactions of Pyrimidine Derivatives with ...

Send Orders of Reprints at [email protected] Current Bioactive Compounds 2013, 9, 000-000

1

Interactions of Pyrimidine Derivatives with Dihydrofolate Reductase and Thymidylate Synthase: Directions Toward Combating toxoplasmosis Letícia C. Assis, Letícia Santos-Garcia, Teodorico C. Ramalho and Elaine F. F. da Cunha† Department of Chemistry, Federal University of Lavras, 37200-000 Lavras, Minas Gerais, Brazil Abstract: Toxoplasma gondii (T. gondii) is the most common cause of secondary central nervous system infection in immunocompromised persons such as AIDS patients. Dihydrofolate reductase and thymidylate synthase enzymes have been studied as attractive targets against parasitic diseases, since they are involved in cell proliferation and influence on DNA synthesis. In this paper, we propose three-dimensional structures of T. gondii dihydrofolate reductase and thymidylate synthase based on homology modeling. In addition, we assessed the interaction mode of pyrimidine analogs in the active site of T. gondii and human enzymes, in order to direct the planning of new compounds that can be used against toxoplasmosis. According to the docking studies, predicted pIC50 values for proposed compounds were higher than those of the experimentally most active compound.

Keywords: Molecular modeling, toxoplasmosis, dihydrofolate reductase and thymidylate synthase. INTRODUCTION The parasitic protozoon Toxoplasma gondii (T. gondii) is the etiologic agent of toxoplasmosis, a parasitic disease widespread among various warm-blooded animals, including humans [1]. Toxoplasmosis is known to be one of the most prevalent parasitic infections of the central nervous system and causes lethal encephalitis in immunocompromised patients, such as those with acquired immunodeficiency syndrome (AIDS) [2]. In spite of the tragic consequences of toxoplasmosis, the therapy of the disease has not changed in the last 20 years [1]. The current treatment consists of combinations of drugs, such as Pyrimethamine (Daraprim®), trimethoprim-sulfamethoxazole (Bactrin®), Sulfadiazine (Triglobe®) or Clindamycin (Dalacin®) [3]. Newborns with congenital toxoplasmosis are treated for at least a year with a combination of antibiotics. If a woman develops toxoplasmosis during pregnancy, her doctor may prescribe medications that reduce the risk of children developing congenital toxoplasmosis. These drugs include spiramycin (Rovamicina® Periodontil®), Pyrimethamine and Sulfadiazine. To decrease the possibility of developing congenital problems related to the drugs, the type and duration of treatment will depend on which quarter she is pregnant [4]. T. gondii thymidylate synthase-dihydrofolate reductase (TS-DHFR) is an essential enzyme in nucleotide biosynthesis and a validated molecular drug target in toxoplasmosis [5]. In humans, TS and DHFR are two separate proteins. In T. gondii, however, TS-DHFR is bifunctional, with both TS and DHFR active sites on a single polypeptide chain of the enzyme. In a previous work, Popov et al created a homology model of TgDHFR based on the closely related structure of DHFR from Mus musculus (PDB ID: 1U70) [6]. However, †

Address correspondence to this author at the Department of Chemistry, Federal University of Lavras, 37200-000 Lavras, Minas Gerais, Brazil; Tel: 55 35 3829-1522; Fax: 55 35 3829-1271; E-mail: [email protected] 1573-4072/13 $58.00+.00

there is no experimentally determined structure of TgTS. Several TS and DHFR inhibitors, as separate entities, have found clinical utility as antitoxoplasmosis agents. Recently, nonclassical analogus were synthesized by Gangjee et al as potential dual thymidylate synthase (TS) and dihydrofolate reductase (DHFR) inhibitors [6]. We have had a longstanding interest in understanding the binding mode of these compounds inside humans (HsTS and HsDHFR) and T. gondii (TgTS and TgDHFR). Thus, we conducted molecular modeling studies involving molecular homology and docking of a number of classical and nonclassical TS-DHFR inhibitors. METHODS Sequence Alignment The primary structures refer to amino acid sequences of TgTS and TgDHFR that were retrieved from the UniProt Knowledgebase [7]. This program manipulates and compares proteins and sequences for homology modeling. As a first approach, an extensive search of potential templates for TgTS and TgDHFR was carried out using the Basic Local Alignment Search Tool (BLAST), a set of programs designed to perform similarity searches [8]. The search for the best template for modeling was carried out by choosing structures possessing a high degree of sequence similarity for each protein, TgTS and TgDHFR. The crystal structural coordinates of Homo sapiens TG (PDB code: 1JU6) [9] was used as a template structure to build a three-dimensional model of TgTS. Sequence alignment was performed using ClustalW [http://www.ch.embnet.org/software/ClustalW. html] [10] and then adjusted according to Delfino et al [23]. In agreement with Popov et al [6] the crystal structural coordinate of Mus musculus DHFR (PDB antry1U70, resolution=2.33Å) [9] was used as a template structure to build a three-dimensional model of TgDHFR.

© 2013 Bentham Science Publishers

2 Current Bioactive Compounds 2013, Vol. 9, No. 2

Table 1.

Assis et al.

Structure of the 6-methyl and 6-ethyl-2-amino-4-oxo-5-substituted thieno[2,3-d]pyrimidines [6]. Compound

Structure O O O

1

O

NH

S

HN

S

N

H 2N

O O O

2

O

NH

S

HN

H 2N

S

N

O S

3

HN

N

H 2N

S

R= Ph

Cl O S

4

HN

N

H 2N

S

4-Cl-Ph

NO2 O S

5

HN

H 2N

N

S

4-NO2-Ph

Interactions of Pyrimidine Derivatives with Dihydrofolate Reductase

Current Bioactive Compounds 2013, Vol. 9, No. 2

3

Table (1) contd…. Structure

Compound

H3C O

O

O

S

6

CH3

HN

S

N

H 2N

2,5-di-OMe-Ph Cl

Cl O S

7 HN

S

N

H 2N

3,4-di-Cl-Ph Cl

S

Cl

O

8 HN S N H2N

3,5-Cl-Ph

9

S O

HN S N H2N

2-Naphthyl


Assis et al.

Table (1) contd….

Compound

Structure N

S O

10 HN S N H2N

Pyridin-4-yl F

S O

11 HN S N H2N

4-F-Ph Br

S O

12 HN S N H2N

4-Br-Ph Cl

S O

13

HN S N H2N

3-Cl-Ph



5

Table (1) contd….

Compound

Structure O

S

O

O

14

HN S N H2N

3,5,di-OMe-Ph Cl

S O

15 HN S N H2N

2-Cl-Ph

Model Building, Refinement and Evaluation

Ligands Data Set

An automated homology model was performed using the SWISS Model server, an automatic web server for protein molecular modeling [11,7]. After identifying homologous proteins and performing the alignment of both template and target sequences, it extracted geometrical restraints (dihedral angles and distances) for corresponding atoms between the query and the template and performed the three-dimensional construction of the protein by using a distance geometry approach. Hydrogen atoms were added to the modeled structure and the atomic coordinates were minimized by the use of the following the protocol according to da Cunha et al [12], to which constraints and restraints were added in order to gain better control over structure relaxation.

The compounds for analysis Table 1 were selected from a series of 6-methyl and 6-ethyl-2-amino-4-oxo-5-substituted thieno[2,3-d]pyrimidines as a potent dual inhibitor of TS and DHFR with IC50 (half maximal inhibitory concentration) values in the μM range. The 3D structure for each compound in their neutral forms was constructed using the Spartan software [17,18], each structure was geometry-optimized without any restriction in vacuum, and the partial atomic charges were assigned using AM1 semi-empirical method [19,20,21].

2'-Deoxyuridine 5'-monophosphate (dUMP) was added to the TgTS previously aligned with 1JU6 and nicotinamide adenine dinucleotide phosphate (NADP) was added to the TgDHFR previously aligned with 7U70. The validation of the generated model was done with the PROCHECK [13] and WHATIF [14] available in the Biotech Validation Suite for Protein Structure [http://biotech.ebi. ac.uk:8400/] [15]. The superposition between the template backbone and the target enzymes was available using the Swiss-Pdb Viewer 3.7 program [16].

Docking Studies The compounds of Table 1 were docked into the TgTS, HsTS, TgDHFR and HsDHFR binding site using the Molegro Virtual Docker (MVD) [22], a program for predicting the most likely conformation of how a ligand will bind to a macromolecule. Ligand and protein are considered flexible during the docking simulation. The MolDock scoring function used by MVD is derived from the piecewise linear potential (PLP), a simplified potential whose parameters are fit to protein–ligand structures and binding data scoring functions and further extended in Generic Evolutionary Method for molecular DOCK with a new hydrogen bonding term and new charge schemes. The docking scoring function, Escore, is defined by the following energy terms:


Assis et al.

Escore = Einter + Eintra

(1)

where Einter is the ligand–protein interaction energy: Einter =

iligand

jprotein

EPLP rij + 332.0 qiqj 4rij2

( )

(2)

The EPLP term is a ‘piecewise linear potential’ using two different sets of parameters: one set for approximating the steric (van derWaals) term between atoms, and another stronger potential for hydrogen bonds. The second term describes the electrostatic interactions between charged atoms. It is a Coulomb potential with a distance-dependent dielectric constant given by: D(r) = 4r. The numerical value of 332.0 adjusts the units of the electrostatic energy to kilocalories per mole [22], Eintra is the internal energy of the ligand:

Eintra =

iligand

jligand

EPLP(rij ) +

flexiblebonds

The first term (double summation) is between all atom pairs in the ligand excluding atom pairs which are connected by two bonds. The second term is a torsional energy term, where u is the torsional angle of the bond. The average of the torsional energy bond contribution is used if several torsions could be determined. The last term, Eclash, assigns a penalty of 1000 if the distance between two heavy atoms (more than two bonds apart) is less than 2.0 Å, ignoring unfeasible ligand conformations [30]. In summary, these functions are used to automatically superimpose a flexible molecule onto a rigid template molecule. The docking search algorithm used in MVD is based on an evolutionary algorithm, where interactive optimization techniques inspired by Darwinian evolution theory, and a new hybrid search algorithm called guided differential evolution. The guided differential evolution algorithm combines the differential evolution optimization tech-

A[1 cos(m. 0 )] + Eclash

Fig. (1). Alignment between TgTS and sequence of the 1JU6 available in PDB.

(3)



7

Fig. (2). Alignment between TgDHFR and sequence of the 1U70 available in PDB.

nique with a cavity prediction algorithm during the search process, which allows for a fast and accurate identification of potential binding modes (poses). The active site exploited in docking studies was defined, with a subset region of 10.0 Å from the centre of the ligand. The interaction modes of the ligand with the enzyme active site were determined as the highest energy scored protein–ligand complex used during docking [23,24]. RESULTS AND DISCUSSION Primary Sequence Comparison and Quality of the 3D Model As a first step, the alignment of the TgTS and TgDHFR primary structure with the chosen template sequences was performed. Figs. (1 and 2) showed the final alignments, which were submitted to the SWISS Model server [11,7] to build the initial models. The alignments of the primary structures between TgTS (residue 322-610) and 1JU6 showed 64.0 % identity, while the alignment between TgDHFR (residue 4-250) and 1U70 showed 38.0 % identity. According to the literature a degree of identity over 30% is satisfactory [25]. The proposed models were validated using the PROCHECK [13] and WHATIF programs [14]. The

Ramachandran plot of the TgTS and TgDHFR satisfied the tests with 77.2% and 72.7% of the residues in the most favored regions, respectively. In addition, 19.9% (TgTS) and 23.8% (TgDHFR) of the residues are in additional allowed regions. The backbones of the TgTS and TgDHFR modeled monomers were superimposed with templates using SPDBViewer [26] in order to calculate the root-mean-square deviation (rmsd) values and to check the structural compatibility between the active site of the homology models with the templates. The superimposition gives rmsd values equal to 1.26 Å, and 1.91 Å for TgTS and TgDHFR subunits, respectively. Binding Site The docked binding mode is used to establish a link between the MolDock scoring function, structural properties of pyrimidine analogs and their biological activity against the TS. The potential binding sites of TgTS and HsTS (protein used in the docking studies) were calculated using Molegro cavity prediction algorithm. A cavity of 283.65 A3 (surface = 871.68 A2) was observed in the TgTS and the amino acid residues from 6 Å of the cavity surface are shown in Table 2. The potential binding site of HsTS (PDB ID: 1HVY) showed a cavity of 260.10 A3 (surface = 637.44 A2), close to amino


Table 2.

Assis et al.

Comparison among the active site residues of site one of TgTS and 1HVY. Residues that do not match are shown in bold.

TgTS

1HVY (HsTS)

TgTS

1HVY (HsTS)

Met91

Lys47

Lys155

Ala111

Asp92

Asp48

Asn156

Asn112

Asp93

Asp49

Phe161

Phe117

Arg94

Arg50

Arg165

Leu121

Thr95

Thr51

Tyr179

Tyr135

Val97

Thr53

Met234

Met190

Val99

Thr55

Ala235

Ala191

Lys121

Lys77

Leu236

Leu192

Arg122

Arg78

Pro237

Pro193

Val123

Val79

His240

His195

Phe124

Phe80

Arg260

Arg215

Gly127

Gly83

Ser261

Ser216

Glu131

Glu87

Asp263

Asp218

Val150

Val106

Leu266

Leu222

Ile152

Ile108

Gln355

Lys308

Trp153

Trp109

acids reported in Table 2. An inspection of Table 2 shows that the TgTS site presents six different residues when compared to the active site of HsTS Met91/Lys47, Val97/Thr53, Val99/Thr55, Lys155/Ala111, Arg165/Leu121 and Gln355/ Lys308. Thus, there is over 81% identity between the enzyme active sites. These amino acids are important for selectivity between TgTS and HsTS. For example, valine has an apolar side chain and exhibits a hydrophobic character, while threonine has a polar side chain and exhibits a hydrophilic character. Thus, an analogue having an apolar group at the position next to those amino acids, will be interacting more strongly with Val in the TgTS and will be repelled by Thr in the HsTS. Among the six different residues, Lys155/Ala111 is closer to the inhibitor. The potential binding sites of TgDHFR and HsDHFR (protein used in the docking studies) were also calculated and the cavities of 496.64 A3 (surface = 1355.52 A2) and 323.584 A3 (surface = 847.36 A2) were observed in the TgDHFR and HsDHFR (3GHY), respectively. The amino acid residues from 6 Å of the cavity surface are shown in Table 3. An inspection of Table 3 shows that the TgDHFR site presents five different

residues when compared to the active sites of HsDHFR: Val8/Ile7, Asp31/Glu30, His34/Tyr33, Ser36/Gln35, Lys40/ Thr39, Met87/Ile60, Pro88/Pro61 and Phe92/Asn64. Thus, there is over 69% identity among the enzyme active sites. These amino acids are important for selectivity between TgDHFR and HsDHFR. Pyrimidines Docking Orientation into Enzymes Through docking studies the following parameters were calculated Table 3b: (a) hydrogen bonding energy between the ligand and the protein (EHBond); (b) interaction energy between the ligand and the cofactor (ELig-Cof); (c) internal energy values of ligand (EIntra); (d) short-range (r4.5Å) electrostatic protein-ligand interaction energy (EElectLong) and (f) steric interaction energy between the protein and the ligand (ESteric). Multiple linear regression (MLR) was used to model the relationship between the independent energy term values and pIC50 values by fitting a linear equation to the observed data. The calculated pIC50 was obtained by the following equation:

pIC50 = C + C1E HBond + C2 E LigCof + C3E Intra + C4 EElect + C5 EElectLong + C6 Esteric

(4)



9

Table 3a. Comparison among the active site residues of site one of TgDHFR and 3GHY. Residues that do not match are shown in bold. TgDHFR

3GHY(HsDHFR)

TgDHFR

3GHY(HsDHFR)

Val8

Ile7

Met87

Ile60

Val9

Val8

Pro88

Pro61

Ala10

Ala9

Lys90

Lys63

Leu23

Leu22

Phe91

Asn64

Trp25

Trp24

Arg92

Arg65

Asp31

Glu30

Pro93

Pro66

Phe32

Phe31

Leu94

Leu67

Lys33

Lys32

Val95

Val68

His34

Tyr33

Arg97

Arg70

Phe35

Phe34

Val151

Val115

Ser36

Gln35

Gly152

Gly116

Lys40

Thr39

Tyr157

Tyr121

Trp52

-

Tyr170

Phe134

Thr83

Thr56

Thr172

Thr136

Ser86

Ser59

Table 3b. Interaction energies for TgTS binding site. Compd

pIC50Exp

pIC50Pred

E HBond

E LigCof

E Intra

EElect

EElectLong

ESteric

EScore

1

7.444

6.938

-5.262

-22.714

0.000

-1.367

-2.446

-65.849

-97.638

2

7.046

7.032

-5.955

-24.124

-3.016

-1.684

-2.092

-32.918

-69.789

3

5.076

5.524

-2.500

-18.348

-4.935

0.000

-4.316

-63.087

-93.186

4

5.658

5.545

-5.066

-22.957

-6.603

-3.623

-0.173

-71.087

-109.509

5

6.000

5.467

-2.500

-9.041

-6.948

-2.062

-1.987

-71.334

-93.873

6

4.638

4.826

-4.937

-35.658

-9.793

-2.701

-0.940

-65.631

-119.661

7

6.000

6.147

-4.630

-28.358

-7.972

-3.322

-0.101

-48.040

-92.423

8

5.796

-

-5.219

-23.956

-8.991

-2.820

-0.839

-70.198

-112.023

9

6.000

6.414

-4.987

-34.711

-1.450

-2.036

-1.684

-76.360

-121.229

10

5.770

6.030

-5.407

-24.046

-4.497

-2.880

-0.930

-65.520

-103.280

11

5.585

5.497

-5.119

-27.901

-6.464

-2.335

-1.458

-65.316

-108.594

12

5.658

5.990

-5.000

-5.494

-7.068

-2.212

-1.252

-61.962

-82.988

13

5.745

5.374

-5.498

-31.185

-7.124

-2.096

-1.516

-66.318

-113.737

14

5.638

5.477

-4.667

-32.538

-9.690

-2.621

-0.678

-56.939

-107.133

15

5.602

5.852

-4.960

-11.552

-6.322

-3.653

0.000

-72.552

-99.039

where C, C1, C2, C3, C4, C5 and C6 are coefficients obtained from the multivariable regression. Timidilate Synthase In the TgTS enzyme, the energy term values of the best docked conformations correlated with inhibition activity,

showed the coefficient of determination (R2) equal to 0.77 (Eq. 5). An R2 greater than 0.7, which indicates a high correlation, was determined between the data [27]. This suggests that the binding conformations and binding models of the TgTS inhibitors are satisfactory. Equation 5 was established with 14 compounds (1-7, 9-15).


Assis et al.

pIC50 = 13.221+ 0.134E HBond + 0.010E LigCof + 0.244E Intra + 0.865EElect + 1.004EElectLong + 0.026ESteric Table 4.

(5)

Interaction energies for the HsTS binding site.

Compd

pIC50Exp

pIC50Pred

E HBond

E LigCof

E Intra

EElect

EElectLong

ESteric

EScore

1

7.398

7.248

-11.649

-33.062

-2.291

-7.481

-5.314

-129.663

-189.460

2

7.268

7.491

-11.063

-33.525

0.000

-8.629

-3.410

-136.696

-193.323

3

5.352

5.465

-3.750

-33.918

-15.667

-0.577

-5.221

-83.961

-143.094

4

6.125

6.126

-3.579

-33.506

-11.237

-1.432

-3.779

-90.938

-144.471

5

6.658

6.375

-3.483

-32.174

-14.317

-1.535

-4.859

-99.498

-155.866

6

5.337

5.528

-4.461

-35.368

-14.803

-0.051

-4.007

-97.017

-155.707

7

6.638

5.880

-3.312

-34.086

-12.648

-1.535

-4.583

-96.239

-152.403

8

6.161

6.020

-3.034

-32.992

-15.118

-1.703

-4.335

-96.148

-153.330

9

5.638

6.006

-3.429

-35.358

-7.524

-2.304

-2.737

-97.370

-148.722

10

6.553

6.434

-3.457

-33.8014

-5.377

-1.233

-4.439

-90.225

-138.532

11

5.886

5.922

-3.582

-33.1921

-14.584

-1.622

-4.107

-91.031

-148.118

12

6.409

6.531

-3.618

-33.432

-7.186

-1.228

-4.142

-94.831

-144.437

13

6.42

6.408

-3.599

-33.995

-5.466

-1.269

-4.882

-93.808

-143.019

14

5.921

6.224

-0.480

-29.722

-21.163

0.000

-8.296

-91.485

-151.146

15

5.959

5.958

0.000

-31.099

-15.924

-4.817

-3.514

-80.140

-135.494

It should be kept in mind that negative energy values represent a higher stability of complexes between the ligand and the protein. EElectLong has great weight in the potency of the compounds since the magnitude of its regression coefficient is 1.000 (Eq. 5). Thus, the positive coefficients correspond to unfavorable interactions between the ligand and the amino acid residues or cofactor in the protein active site, in order words these interactions decrease the potency of the inhibitors. For example, Compound 3 has less inhibitory potency value and showed the least EElectLong Table 4. This is in good agreement with the experimentally observed results. By analyzing the hydrogen bond formed between the 15 compounds and TgTS, it was observed that all compounds interact with Asp263 and cofactor through a pyrimidine ring. Compound 9 exhibited the lowest EScore (more stable complexes). Fig. (3) shows Compound 9 docked into the TgTS active site. It can be observed that the 2-naphthyl substituent interacts with Phe270 and Phe124 through -staking interactions; the pteridine ring interacts with Asp263 and cofactor through hydrogen bonds. The R2 value for the HsTS enzyme was slightly higher than for TgTS, 0.80. Equation 6 was established with 15 compounds (1-15).

Fig. (3). Docking of the conformation corresponding to Compound 9 inside the TgTS active site.

E LigCof E LigCof has great weight in the potency of the compounds since the magnitude of its regression coefficient is 0.300 (Eq. 6). Thus, the positive coefficients correspond to unfavorable interactions. Compound 6, for example, has the least E LigCof E LigCof value and it showed the least inhibitory potency in this series of compounds Table 4. By analyzing the hydrogen bond formed between the 15 compounds and HsTS we observed: (i) all compounds interact with Asp218 (equivalent to Asp263 in TgTS) and cofactor; (ii)

pIC50 = 15.480 0.011E HBond + 0.300E LigCof + 0.077E Intra + 0.082EElect + 0.082EElectLong 0.021ESteric

(6)


only Compounds 1 and 2 interact with Lys77, Leu221; (iii) only Compounds 1, 2 and 5 interact with Lys308. Among the nonclassical compounds, Compound 5 showed more stable interaction energy with TgTS. The compound with an electron withdrawing substituent (4-nitrophenyl) on the side chain phenyl ring was more potent because it is the only nonclassical compound which interacts with Lys308. In the TgTS this amino acid is equivalent to Gln355. Compound 2 showed the lowest EScore (more stable complexes, Table 4) and it is evi-


11

E HBond has great weight in the potency of the compounds since the magnitude of its regression coefficient is 0.172(Eq. 6). Thus, the negative coefficients correspond to favorable interactions. Compounds 1 and 2 are the most potent compounds in the series against HsDHFR and showed the lowest E HBond E HBond value Table 5.

dent from Tables 3 and 4 that the ratio between EScore

EScore of TgTS and HsTS enzymes is about 3.0. The 4-C6H4CO-L-Glu substituent of Compound 2 is located next to polar amino acids (i.e. lysine) in the HsTS active site, while this same substituent is located next to non-polar amino acids (i.e phenylalanina) in the TgTS active site. Fig. (4) shows Compound 2 docked into the HsTS active site. As discussed previously, the TgTS cavity space is larger than the HsTS cavity space, thus, the biological activity is more sensitive to the flexibility of inhibitors with HsTS. In addition, Lys155 is the amino acid in the TgTS next to the sulfur atom of the pteridine ring of the inhibitors. In the HsTS this amino acid is represented by Ala112. Thus, it is still very important to research substitution in the rings in both enzymes. Dihydrofolate Reductase The R2 value for the HsDHFR enzyme was 0.81. Equation 7 was established with 15 compounds (1-15).

Fig. (4). Docking of the conformation corresponding to Compound 2 inside the HsTS active site.

By analyzing the hydrogen bond formed between the 15 compounds and HsDHFR we observed: (i) all compounds interact with Glu30; (ii) only Compounds 1, 2, 5, 16 and 18 interact with Asn64; (iii) only Compounds 1, 2 and 9 interact with Thr136; (iv) only Compounds 1 and 2 interact with Arg70. Among Compounds 3-15, the 9th exhibited the most stable interaction energy with HsDHFR. The hydrogen bond formed between Thr64 and the naftil group can be responsible for the stability of Compound 9. In addition, Compound 9 exhibited -staking interactions with Phe31. Fig. (5) shows Compound 9 docked into the HsDHFR active site.

pIC50 = 6.647 0.172E HBond + 0.119E LigCof + 0.009E Intra + 0.126EElect + 0.163EElectLong 0.012ESteric Table 5.

(7)

Interaction energies for the HsDHFR binding site.

Compd

pIC50Exp

pIC50Pred

E HBond

E LigCof

E Intra

EElect

EElectLong

ESteric

EScore

1

7.699

7.887

-14.354

-14.810

0.000

-5.096

-4.393

-152.366

-190.563

2

7.721

7.416

-12.967

-16.218

-2.541

-5.402

-4.605

-159.515

-201.251

3

5.585

6.012

-5.138

-21.834

-6.395

-0.485

-0.297

-103.528

-137.680

4

6.523

6.523

-5.043

-18.089

-15.171

-0.154

-0.189

-112.169

-150.819

5

6.585

6.582

-6.014

-18.512

-8.853

-0.559

-1.085

-118.845

-153.872

6

6.076

6.143

-3.945

-21.253

-7.963

-0.151

-0.491

-126.009

-159.814

7

6.538

6.395

-4.607

-18.676

-16.458

-0.467

-0.137

-117.03

-157.377

8

6.553

5.909

-4.659

-21.650

-10.585

-0.908

-0.137

-105.532

-143.473

9

5.658

6.047

-5.509

-20.032

-15.025

-2.846

-1.166

-126.444

-171.025

10

6.456

6.421

-4.533

-18.115

-13.229

-0.132

0.000

-106.93

-142.94

11

6.208

6.261

-3.760

-17.840

-14.145

-0.369

-0.497

-111.872

-148.485

12

6.585

6.654

-5.656

-18.193

-14.925

-0.015

-0.068

-112.000

-150.860

13

6.523

6.503

-4.921

-17.9454

-15.628

0.000

-0.209

-109.742

-148.447

14

6.268

6.323

-5.028

-18.442

-17.712

-0.336

-1.035

-114.508

-157.062

15

6.222

6.114

-3.079

-18.803

-11.460

-0.244

-0.143

-110.818

-144.550


Assis et al.

E LigCof (positive coefficient) has great weight in the potency of the compounds since the magnitude of its regression coefficient is 0.113 (Eq. 8). Compound 14 has the lowest value for E LigCof and it showed the least inhibitory potency in this series of compounds. Similar to Equation 6, compounds with methoxy substituent did not contribute to stability of interaction with the cofactor Table 6.

Fig. (5). Docking of the conformation corresponding to Compound 9 inside the HsDHFR active site.

By looking the hydrogen bond formed between the 15 compounds and TgDHFR we observed: (i) all compounds interact with Asp31 and Thr172; (ii) only Compounds 1, 2, 5 and 14 interact with Ser36; (iii) only Compound 2 interacts with Trp52; (iv) only Compound 5 interacts with Arg97. Compound 5 Fig. (6) also showed more stable interaction energy with TgDHFR. It had the lowest energy value for

EElect and E HBond terms and hydrogen bonds between the nitro group and Ser36 and Arg97. The docking studies pointed out that analogs with electron withdrawing substitutions on the phenyl ring were more potent than other analogues, corroborating with experimental data [28]. Correlating the docking results obtained, it is interesting to Fig. (6). Docking of the conformation corresponding to Compound 5 inside TgDHFR active site. 2

The R value for the TgDHFR enzyme was 0.83. Equation 8 was established with 14 compounds (1-3 and 5-15).

note that the HsDHFR EScore

HsTS EScore

TgTS values are higher than the EScore

TgDHFR values in most compounds, in other words, EScore

the pirimidine analogs studied are more selective for the TS

pIC50 = 7.931 0.006E HBond + 0.113E LigCof + 0.034E Intra 0.042EElect + 0.014EElectLong 0.012ESteric Table 6.

(8)

Interaction energies for TgDHFR binding site.

Compd

pIC50Exp

pIC50Pred

E HBond

E LigCof

E Intra

EElect

EElectLong

ESteric

EScore

1

8.097

8.027

-6.261

-12.000

-5.242

-1.999

-2.578

-128.790

-156.871

2

8.678

8.617

-6.614

-9.566

0.000

-2.127

-2.426

-139.375

-160.110

3

7.481

7.700

-6.137

-11.647

-11.117

-2.608

-2.171

-112.301

-145.978

4

8.046

-

-0.753

-27.724

-7.240

-0.123

-4.206

-105.832

-133.807

5

8.060

8.106

-8.466

-11.935

-4.684

-5.179

-2.233

-120.566

-153.066

6

7.854

7.870

-2.208

-11.494

-11.262

-2.445

-2.377

-128.214

-158.002

7

8.092

8.077

-6.158

-10.338

-9.440

-5.275

0.000

-114.74

-145.953

8

7.854

8.018

-6.140

-9.975

-11.930

-4.062

-1.016

-118.956

-152.082

9

8.076

8.246

-4.272

-10.259

-2.643

-0.662

-3.530

-130.076

-151.445

10

7.770

7.816

-2.516

-12.005

-8.301

-4.040

-0.997

-112.796

-140.658

11

7.602

7.600

-6.145

-13.193

-7.745

-2.595

-2.346

-109.183

-141.210

12

7.886

7.693

-5.653

-10.644

-12.444

-2.591

-2.276

-106.486

-140.095

13

7.921

7.750

-2.392

-11.347

-11.8302

-2.501

-2.243

-118.006

-148.324

14

7.620

7.651

-6.757

-13.151

-9.208

-1.904

-2.335

-119.290

-152.647

15

7.569

7.723

-2.359

-11.370

-11.340

-2.454

-2.219

-114.767

-144.511


Table 7.


13

Proposed structures of three new potential inhibitors. TgTS

HsTS

TgDHFR

HsDHFR

pIC50Pred (Eq. 5)

pIC50Pred (Eq 6)

pIC50Pred (Eq. 8)

pIC50Pred (Eq. 7)

9.425

5.536

8.099

7.126

10.111

5.525

8.098

7.504

10.106

7.236

9.427

7.138

O O

S O

HN H2 N

S

N

A COOH

O

S

HN H2 N

S

N

B

O

COOH S

HN N

N

S

C

enzyme. Those analyses lead to the proposition of the structures of the three derivatives as potential inhibitors of TS and DHFR which are shown in Table 7. The proposed compounds were evaluated using docking simulations at the active sites of the TgTS, HsTS, TgDHFR and HsDHFR enzymes. The predicted pIC50 values given in Table 7 for the proposed compounds indicate that the three compounds are more potent for TgTS and TgDHFR. The inhibitory potency of Compound C with TgDHFR was higher than that of the experimentally most active compound. The same was observed among Compounds A, B and C and TgTS. According to the docking studies, B and C are promising TgTS inhibitors.

sible for activity and selectivity of new DHFR and TS inhibitors targeting a potential drug to combat T. gondii infection.

CONCLUSION

[1]

In this work, we have used homology modeling to propose the 3D structure for T. gondii TS and DHFR. Our molecular docking results suggest that differences between Lys155/Ala111 in TgTS/HsTS and Asp31/Glu30 in TgDHFR/HsDHFR seem to afford the only good opportunities to be exploited in the design of compounds that would selectively bind to TgTS and TgDHFR enzymes. Analyses of the interactions of the known pyrimidine analogs with the active sites of these models (TgTS and TgDHFR) and human TS and DHFR were performed. Good and reliable correlations (R2>0.7) between the energy terms and pIC50 were obtained. Three compounds were proposed and through docking studies we observed that B and C were more potent against TgTS. This study may be helpful in understanding the molecular interactions and the structural factors respon-

CONFLICT OF INTEREST The authors confirm that this article content has no conflicts of interest. ACKNOWLEDGEMENTS We thank the Brazilian agency FAPEMIG, CNPq for funding part of this work. REFERENCES

[2]

[3]

[4] [5] [6]

El Kouni, M.H.; Guarcello, V.; Al Safarjalani, O.N.; Naguib, F.N.M. (1999) Metabolism and selective toxicity of 6nitrobenzylthioinosine in Toxoplasma gondii. Antimicrob. Agents Chemother., 43, 2437. Al Safarjalani, O.N.; Rais, R.H.; Ah Kim, Y.; Chu, C.K.; Naguib, F.N.M.; el Kouni, M.H. (2007) 7-Deaza-6-benzylthioinosine analogues as subversive substrate of Toxoplasma gondii adenosine kinase: Activities and selective toxicities. Biochem. Pharmacol., 73,1558 da Cunha, E.F.F.; Ramalho, T.C.; Mancini, D. T.; Fonseca, E.M.B.; Oliveira, A.A. (2010) New Approaches to the Development of Anti-Protozoan Drug Candidates: a Review of Patents J. Braz. Chem. Soc., 21, 1787. Petersen, E. (2007) Prevention and treatment of congenital toxoplasmosis. Expert. Rev. Anti. Infect. Ther., 5, 285. Reynolds, M.G.; Roos, D.S. (1998) A biochemical and genetic model for parasite resistance to antifolates. J. Biol. Chem., 273, 3461. Popov, V. M.; Yee, W.A.; Anderson, A.C. (2007) Towards in silico lead optimization: Scores from ensembles of protein/ligand con-


[7] [8] [9]

[10]

[11] [12]

[13] [14] [15] [16] [17] [18]

Assis et al.

formations reliably correlate with biological activity. Proteins: Struct., Funct., Bioinf., 66, 375. Peitsch, M. C. (1996) ProMod and Swiss-Model: Internet-based tools for automated. Biochem. Soc. Trans., 24, 274. Altschul, S.F.; Gish, W.; Miller,W.; Myers, E.W.; Lipman,D.J. (1990) Basic local alignment search tool. J. Mol. Biol., 215, 403. Sayre, P.H.; Finer-Moore, J.S.; Fritz, T.A.; Biermann, D.; Gates, S.B.; MacKellar, W.C.; Patel, V.F.; Stroud, R.M. (2001) Multitargeted antifolates aimed at avoiding drug resistance form covalent closed inhibitory complexes with human and Escherichia coli thymidylate synthases. J. Mol. Biol., 313, 813. Thompson, J.D.; Higgins, D.G.; Gibson, T.J. (1994) CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res., 22, 4673. Peitsch, M.C. (1995) Protein modeling by mail Biotechnology. Biochem. Soc. Trans., 13, 658. da Cunha, E.F.F.; Ramalho, T.C.; Alencastro, B. R.; Maia, E.R. (2004) Interactions of 5-deazapteridine derivatives with Mycobacterium tuberculosis and with human dihydrofolate reductases. J. Biomol. Struct. Dyn., 22, 119. Laskowski, R. A.; Moss, D. S.; Thornton, J. M. (1993) Main-chain bond lengths and bond angles in protein structures. J. Mol. Biol., 231, 1049. Vriend, G. (1990) What if – a molecular modeling and drug design program, J. Mol. Graph., 8, 52. Laskowski, R.A. (2001): PDBsum: summaries and analyses of PDB structures. Nucleic Acids Res., 29, 221. Guex, N.; Diemand, A.; Peitsch, M.C.; Schwede, T. GlaxoSmithKline R&D & the Swiss Institute of Bioinformatics. SPARTAN, version 10.0; Wavefunction, Inc.: Irvine, CA. Young, D. C. Computational Chemistry: A Practical Guide for Applying Techniques to Real-World Problems. WileyInterscience., 2001.

Received: October 23, 2012

[19]

[20] [21]

[22] [23] [24]

[25] [26]

[27]

[28]

Revised: April 19, 2013

Cramer,C.J.; Truhlar, D.G. (1991) General parameterized SCF model for free energies of solvation in aqueous solution. J. Am. Chem. Soc., 113, 8305. Cramer,C.J.; Truhlar,D.G. (1992) An SCF solvation model for the hydrophobic effect and absolute free energies of aqueous solvation. Science, 256, 213. Cramer,C.J.; Truhlar,D.G. (1992) AM1-SM2 and PM3-SM3 parameterized SCF solvation models for free energies in aqueous solution. J. Comput.-Aided Mol. Des., 6, 629. Thomsen, R.; Christensen, M.H. (2006) MolDock: A new technique for high-accuracy molecular docking. J. Med.Chem., 49, 3315. Trifono, E.N. (2011) Vocabulary of definitions of life suggests a definition. J. Biomol. Struct. Dyn., 29, 259. Rooman, M.; Cauët, E.; Liévin, J.;Wintjens,R. (2011) Conformations consistent with charge migration observed in DNA and RNA x-ray structures. J. Biomol. Struct. Dyn., 28, 949. Brenner, S.E.; Chothia, C.; Hubbard, T.J.P. (1998) Assessing sequence comparison methods with reliable structurally identified distant evolutionary relationships. Proc. Natl. Acad. Sci., 95, 6073. Altschul, S. F.; Madden, T. L.; Schaffer, A. A.; Zhang, J.; Zhang, Z.; Mille, W.; Lipman, D. (1997) Gapped BLAST and PSIBLAST: a new generation of protein database search programs. Nucleic Acids Res., 25, 3389. Hopfinger,A.J.; Wang,S.; Tokarski,J.S.; Jin, B.; Albuquerque, M.; Madhav,P.J.; Duraiswami, C. (1997) Construction of 3D-QSAR models using the 4D-QSAR analysis formalism. J. Am. Chem. Soc., 119, 10509. Gangjee, A.; Li, W.; Kisliuk, R.L.; Cody, V.; Pace, J.; Piraino, J.; Makin, J. (2009) Design, synthesis, and x-ray crystal structure of classical and nonclassical 2-amino-4-oxo-5-substituted-6ethylthieno[2,3-d]pyrimidines as dual thymidylate synthase and dihydrofolate reductase inhibitors and as potential antitumor agents. J. Med. Chem., 52, 4892.

Accepted: April 22, 2013