design of new uracil derivatives possessing inhibitory ... - Springer Link

3 downloads 0 Views 85KB Size Report
DESIGN OF NEW URACIL DERIVATIVES POSSESSING. INHIBITORY ACTIVITY WITH RESPECT TO REVERSE. TRANSCRIPTASE OF HIV-1 MUTANT K103N/ ...
DOI 10.1007/s11094-016-1353-x Pharmaceutical Chemistry Journal, Vol. 49, No. 10, January, 2016 (Russian Original Vol. 49, No. 10, November, 2015)

DESIGN OF NEW URACIL DERIVATIVES POSSESSING INHIBITORY ACTIVITY WITH RESPECT TO REVERSE TRANSCRIPTASE OF HIV-1 MUTANT K103N/Y181C S. V. Pechinskii,1 A. G. Kuregyan,1 A. A. Ozerov,2 and M. S. Novikov2 Translated from Khimiko-Farmatsevticheskii Zhurnal, Vol. 49, No. 10, pp. 40 – 43, October, 2015.

Original article submitted July 22, 2014. New highly active N1-substituted uracil derivatives of the HIV-1 nonnucleoside reverse transcriptase inhibitor class were designed. The structure—activity relationship provided a basis for building a computer model of the inhibitory activity against reverse transcriptase of HIV-1 mutant K103N/Y181C. New compounds that were more active than nevirapine were synthesized. Keywords: anti-HIV-1 activity, mutant strains, nonnucleoside reverse transcriptase inhibitors, uracil derivatives, molecular docking.

HIV treatment remains one of the most challenging modern medical problems despite enormous efforts of the global scientific community. Over 35 million people in the world, including 700,000 in Russia [1], are infected with HIV-1 and require highly active antiretroviral therapy (HAART), which can be provided by six drug classes, i.e., nucleoside (nucleotide) (NRTI) and nonnucleoside (NNRTI) reverse transcriptase inhibitors in addition to protease, integrase, and fusion inhibitors and CCR5 receptor antagonists [2, 3]. The goal of the present research was to design new highly active NNRTI against HIV-1 mutant K103N/Y181C. An advantage of NNRTI is their high antiviral activity. However, an important drawback is the rapid development of resistant mutants due to the low accuracy of viral DNA and RNA reverse transcriptase. Mutations lead to changes, among others, at the NNRTI binding site. These reduce the efficacy and sometimes even generate resistance [4]. One of the most common and clinically significant mutations is the HIV-1 double mutation K103N/Y181C that results from the use of NNRTI. This reduces by several times the efficacy of drugs such as nevirapine and efavirenz. This mutation is characterized by the substitution of 103 asparagine by lysine and 181 cysteine by tyrosine and alters the configuration of the NNRTI binding site [5]. 1 2

Molecular docking is one of the most effective methods for computer modeling of the interaction of a ligand with a protein. It can estimate quantitatively the strength of such an interaction. We used the ArgusLab GA Dock program (authored by Mark Thompson of Planaria Software, Seattle, WA) [6]. The steric ligand structure was optimized by the PM3 semi-empirical method in the GAMESS program [7]. This combination provided the most accurate modeling of the ligand–protein interaction. A computer model for the interaction of an experimental compound with a structure close to those of the synthesized compounds with reverse transcriptase of double mutant K103N/Y181C that was constructed using a crystal structure at 2.9 Å resolution (PDB code 4H4O) provided a starting point [8]. The obtained predictions were compared with in vitro screening results for a series of uracil derivatives in cell cultures of human MT-4 T-lymphocytes infected with HIV-1 double mutant K103N/Y181C [9]. EXPERIMENTAL CHEMICAL PART PMR and 13C NMR spectra were recorded in DMSO-d6 with TMS internal standard on a Bruker AMXIII-400 spectrometer. Spectra were interpreted using the ACD/HNMR PredictorPro 3.0 licensed program (Advanced Chemistry Development, Canada). TLC analysis was performed on Sorbfil plates using i-PrOH eluent and I2 vapor detector.

Pyatigorsk Medico-Pharmaceutical Institute, Branch of Volgograd State Medical University, MH RF, Pyatigorsk, 357532 Russia. Volgograd State Medical University, MH RF, Volgograd, 400131 Russia.

683 0091-150X/15/4910-0683 © 2015 Springer Science+Business Media New York

684

S. V. Pechinskii et al. Tyr188A

RESULTS AND DISCUSSION

O OH

Cl Cl Cl Tyr181A

O O

N

O

N

R

H

N

O

O Lys103A Fig. 1. Positioning of XVI in active center of reverse transcriptase.

Melting points were measured in glass capillaries on a Mel-Temp 3.0 apparatus (Laboratory Devices Inc., USA). General method for synthesizing uracil derivatives A 2,4-Di(trimethylsilyloxy)pyrimidine (20 – 25 mmol), which was prepared from uracil, thymine, or 6-methyluracil by refluxing in an excess of hexamethyldisilazane, was heated with an equal amount of the appropriate 1-bromo-2(2-benzoylphenoxy)ethane or 1-bromo-2-(2-benzylphenoxy)ethane at 160 – 170°C for 2 h, cooled, treated with a mixture of EtOAc (40 mL) and i-PrOH (10 mL), stirred for 30 min at room temperature, filtered, and evaporated. The solid was recrystallized from i-PrOH or its mixture with DMF (1:1) to afford I-XIX (Table 1). The product yields were 62 – 83% [10]. EXPERIMENTAL BIOLOGICAL PART Anti-HIV-1 activity was studied in vitro using a suspension of CEM-SS cells (105 cells/mL) in growth medium that were infected with HIV-1 double mutant strain K103N/Y181C. Solutions containing various concentrations of the synthesized compounds in DMSO were added quickly after the infection. The samples were incubated for 4 d at 37°C. The number of living cells was found after the incubation was finished using 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide. The concentrations at which CEM-SS cells were 50% protected from the cytopathic effect of HIV-1 (EC50) were determined. The accuracy of the assay was >10% [11].

The majority of clinically significant mutations alters the amino-acid sequence at the active binding site. Therefore, the flexibility, i.e., the ability of the molecular structural fragments to wiggle and shift, must be considered in designing new compounds. Such a “strategic flexibility” concept allows potentially active compounds against a large number of mutations to be sought. Highly active structures were sought among N1-substituted uracil derivatives, which demonstrated previously pronounced activity against an HIV-1 wild strain and various mutants [9, 11]. A training set of compounds that were used to adjust the model was studied in the first stage. The results for their in vitro anti-HIV-1 activity were used to check the accuracy of the built computer model describing the ligand—protein interaction. It was found during building of the computer model that inhibitor binding to lysine 103 was most significant and occurred because of the uracil in the structure. Modeling of the ligand–protein interaction showed that not only the ligand flexibility but also the protein flexibility had to be considered because it played a significant role. It was found that the protein structure adapted to the ligand and could adopt a complementary shape with the borders of the binding pocket surrounding the inhibitor. The local changes in the protein structure could explain the interaction of the ligand with tyrosine 188 and compensated for the loss of the hydrophobic interaction due to the Y181C mutation. This interaction was a second key factor in the ligand–protein binding. The prediction accuracy could be increased considerably and a strategy for designing new highly active compounds of this class could be developed by introducing the flexibilities of both the ligand and protein into the model. Figure 1 shows the positioning of XVI in the reverse transcriptase active site. The model based on relationships between the inhibitor activity and docking energy did not always allow an adequate prediction to be made because even a small change of the inhibitor molecular structure sometimes led to a complete loss of activity. Therefore, use of a multi-factor scoring function was proposed for the prediction. It included estimates of not only the docking energy but also a larger number of factors such as the length and number of H-bonds, ligand lipophilicity, deviation of the N1-substituent from the uracil plane, and flexibility of the ligand–protein interaction that were estimated from the change of bond length near the interaction (Fig. 1). The multi-factor scoring function was calculated using the formula: S = k 1F 1 + k 2F 2 + … + k nF n, where k is the value of the coefficient indicating the extent of influence on the ligand–protein interaction and F, a factor influencing the ligand–protein interaction. The factors were ranked using the experimental results. Each factor was assigned a maximum ball value depending on its influence on the ligand–protein binding. The extent of

Design of New Uracil Derivatives

O

H N

685

O N

H N

O

O N

O

R1

O

O

R2 R4

R4 R3

R3

I – VI

VII – XIX

Fig. 2. Structural formulas of synthesized N1-substituted uracil derivatives.

the influence of the factors and its expression in balls were established by analyzing the training set, which included compounds I-XIV (Table 1). The accuracy of the predictions from the developed model was increased substantially by using this approach to calculate the extent of influence of the various factors on the ligand inhibitor properties. Compounds XV-XIX were studied based on the predictions from the developed model (Table 1, Fig. 2). The reference drugs were nevirapine and efavirenz.

The structure–activity relationship of the various candidates was studied using computer modeling in order to discover new highly active NNRTI. Understanding of how structural changes affected the ability of the compounds to inhibit reverse transcriptase assisted the design of new highly active compounds of this class. The results showed that the accuracy of the prediction was increased if the multi-factor scoring function rather than the docking energy was used to predict the activity. For ex-

TABLE 1. Activity in vitro and in silico of N1-Substituted Uracil Derivatives Against HIV-1 Mutant K103N/Y181C Compound

R1

R2

R3

R4

EC50, mM

Docking energy, kcal/mol

Number of balls, multi-factor scoring function

I

H

H

Me

H







II

H

H

Me

3-Me







III

H

H

Me

3.5-Me2







IV

H

H

Cl

3.5-Me2

> 200

— 10.314

224

V

H

H

Br

3.5-Me2



— 8.235



VI

H

H

Me

H

> 357

— 9.016

82

VII

H

H

F

H

171

— 11.420

301

VIII

H

H

Cl

H

> 337

— 10.642

96

IX

H

H

Br

H

> 301

— 9.859

154

X

H

H

Cl

3-Me

52

— 13.536

877

XI

H

H

Me

3.5-Me2

> 179

— 10.899

278

XII

H

H

Cl

3.5-Me2

> 313

— 9.802

103

XIII

H

H

Br

3.5-Me2

> 282

— 9.011

178

XIV

H

H

Br

3.5-Me2

24.8

— 14.372

1564

XV

H

H

Cl

3.5-F2

> 307

— 14.840

132

XVI

H

H

Cl

3.5-Cl2

1.13

— 16.537

11516

XVII

H

H

Cl

3.5-Br2

> 237

— 13.204

206

XVIII

Me

H

Cl

3.5-Me2

4.09

— 15.155

8018

XIX

H

Me

Cl

3.5-Me2

4.09

— 15.206

8102

Nevirapine









> 15

— 14.023

2097

Efavirenz









0.56

— 17.227

13388

686

ample, calculations in the second stage using the docking energy of XV-XIX did not predict the sharp activity decreases for XV and XVII whereas the predictions were very close to the in vitro results if the multi-factor scoring function was used. The computer model based on the multi-factor scoring function enabled the inhibitor properties of still unsynthesized compounds to be predicted highly accurately. The synthesis of these compounds and in vitro tests confirmed the model predictions. The virtual screening approach of still unsynthesized compounds shortened substantially the time for discovering highly active compounds and also broadened the understanding of the structure–activity relationship. This approach is especially crucial to the development of antiviral drugs because of the vigorous generation of resistant strains. REFERENCES 1. HIV / AIDS; http: // www.who.int / topics / hiv aids / en /

S. V. Pechinskii et al.

2. T. L. Hartman and R. W. Buckheit, Mol. Biol. Int., 1 – 17 (2012). 3. D. Li, P. Zhan, E. De Clercq, and X. Liu, J. Med. Chem., 55, 3595 – 3613 (2012). 4. J. E. Gallant, Int. AIDS Soc. USA Top. HIV Med., 13(5), 138 – 142 (2005). 5. K. Das, J. D. Bauman, A. D. Clark, Jr., et al., Proc. Natl. Acad. Sci. USA, 105(5), 1466 – 1471 (2008). 6. ArgusLab, GA Dock [electronic resource]; http: // www.arguslab.com / arguslab.com / ArgusLab.html 7. GAMESS [electronic resource]; http: // www.msg.chem. Iastate.edu / gamess / download.html 8. K. M. Frey, M. Bollini, A. C. Mislak, et al., J. Am. Chem. Soc., 134, 19501 – 19503 (2012). 9. M. S. Novikov, O. N. Ivanova, A. V. Ivanov, et al., Bioorg. Med. Chem., 19, 5794 – 5902 (2011). 10. A. A. Ozerov, M. S. Novikov, A. I. Luganchenko, et al., Volgogr. Nauchno-Med. Zh., No. 4, 15 – 18 (2012). 11. R. W. Buckheit, E. L. White, V. Fliakas-Boltz, et al., Antimicrob. Agents Chemother., 43, 1827 – 1834 (1999).