In silico Prediction of Pyrazoline Derivatives as Antimalarial agents

 

Sonal Dubey1*, Sakshi Bhardwaj2, Prabitha Parbhakaran3, Ekta Singh4

1College of Pharmaceutical Sc, Dayanand Sagar University, Kumarswamy Layout, Bengaluru – 560078 – India.

2Krupanidhi College of Pharmacy, Carmelaram, Bengaluru - 560035. India.

3JSS College of Pharmacy, SS Nagar Bannimantap, Mysuru - 570015. India.

4Acharya and BM Reddy College of Pharmacy, Soldevanahalli, Bengaluru – 560107 India.

*Corresponding Author E-mail: drsonaldubey@gmail.com

 

ABSTRACT:

Malaria is one of the toughest health and development challenges faced by tropical countries. The resistance of malarial parasite to available drugs and currently used chemotherapy made its emergence for development of new drugs. Pyrazoline derivatives have shown good antimalarial activity. In present work, our objective is to explore pyrazoline derivatives with in silico methods for their antimalarial activity. A five-point pharmacophore was developed using 80 molecules having logIC50 ranging from 10.39 to 6.72. The pharmacophore yielded a statistically significant 3D-QSAR model with a high correlation coefficient R2= 0.806772, cross validation coefficient Q2= 0.7154 at four component PLS factor. To evaluate the effectiveness of docking protocol, we have selected crystallographic bound compound to validate our docking procedure. Protein selected for our studies with PDB id is 2BMA having resolution 2.7 Å. Further similar orientations were observed between the superpositions of 80 compounds after pharmacophore and 3D-QSAR poses, pharmacophore and XP docking poses, 3D-QSAR and XP docking poses. These present studies will provide insight in designing novel molecules with better antimalarial activity. Results explained that two aromatic rings and two hydrophobic groups are important for the antimalarial activity. The docking studies of all selected inhibitors in the active site of 2BMA showed crucial hydrogen bond interactions with HIS95, SER97, GLN323, ARG93, ALA321, ALA346, ILE166, ILE102 and PRO96 amino acid residues.

 

KEYWORDS: Antimalarial, pyrazoline, In silico, QSAR, pharmacophore.

 

 


INTRODUCTION:

Malaria is most fatal disease caused by Plasmodium falciparum. Other species of plasmodium which causes malaria are Plasmodium vivax, P. malariae, P. ovale and P. knowlesi1. A number of drugs and drug combinations are available to treat malaria but many of them become ineffective because of parasite resistance. The major problem associated with antimalarial drugs is development of drug resistant which emergence the need to develop new antimalarial drugs for prevention of drug resistance issue2.

 

According to the report, there were 212 million new cases of malaria worldwide in 2015 (range 148–304 million). The WHO African Region accounted for most global cases of malaria (90%), followed by the South-East Asia Region (7%) and the Eastern Mediterranean Region (2%)3. Pyrazolines are nitrogen-containing heterocyclic compounds, well known for their pronounced biological activity. Pyrazoline compounds have been reported to possess versatile biological activities including anticancer4-5, anti-inflammatory6-7, antiviral8-9, antimicrobial10-11, antimalarial12-14 and anti-leishmanial activities15-17. Chloroquine and other quinine derivatives have been the most commonly used drugs to treat malaria but parasites have developed widespread resistance against these drugs18. To overcome parasite resistant problem, artemisinin-based combination therapy (ACT) has been developed as a first line treatmnet19-20 but after that also parasite developed resistant against this therapy. Hence development of antimalarial drugs is in emergence. In this light, our work contributing the In silico predictions of 80 reported pyrazoline compounds for development of novel pyrazoline analogues with better antimalarial activity against parasite Plasmodium falciparum. The first stage of infection involves the entry of merozoites into erythrocytes, here merozoites released from the matured liver stage parasites into erythrocytes, where they grow into trophozoites, which upon maturation undergo schizogony to form merozoites. At the end of 48 h intraerythrocytic life cycle of parasites, the merozoites released from infected erythrocytes invade new erythrocytes to start the next 48 h life cycle21-22. Parasite depends on host nutrition for its development. Degradation of host cell haemoglobin by aminopeptidases results in formation of amino acids which are the main source for parasite protein synthesis23-24. Recently, computational methods like pharmacophore modelling, 3D-QSAR, docking and molecular dynamics simulation were used successfully by many researchers to discover novel high affinity compounds to treat different diseases. The pharmacophore model of these inhibitors mainly consists of important features:  hydrogen bond donor, hydrogen bond acceptors, one aromatic ring, hydrophobic ring. To gain insights into the structural and chemical features required for antimalarial activity, we developed a ligand-based 3D-pharmacophore hypothesis using PHASE (Schrödinger 2016-4). The pharmacophore hypothesis obtained from the pharmacophoric features was used to derive a pharmacophore based 3D-QSAR model. Then molecular docking studies of the data set molecules was performed using Glide.

 

MATERIALS AND METHODS:

Data Set:

A total of 80 pyrazoline derivatives were selected from literature for pharmacophore, QSAR and docking studies which have reported their activity against Plasmodium falciparum. All the structures had the same pharmacophore with variable substitutions which are contributing difference in the observed antimalarial activity. The logIC50 of these 80 compounds ranging from 10.39 to 6.7225-28. Data set molecules shares a common assay procedure. All the given IC50 values of data set molecules reported in microgram per millilitre which were converted into their respective logIC50 values. The molecules in training set provided their structure feature template to build a QSAR model. Whereas structure feature of test set used for internal validation. Training set and test set molecules are structurally diverse molecules with a difference of log 3 in their reported activity. Energy optimisation and conformation generation of molecules was done by using Ligprep module of Schrodinger 2016 suite.

 

Pharmacophore hypothesis generation and validation:

The pharmacophore hypothesis for given data set of molecules was generated using Phase module of Schrodinger 2016 suite. To correlate biological activity with 3D descriptors partial least square (PLS) approach was used. An activity threshold was given such that the molecules with logIC50 greater than 9.30 were considered as active and below than 7.30 were considered as inactive.  While rest of the molecules were considered moderately active. Pharmacophoric sites were generated with the help of structural features of active molecules. All the scores for actives and inactives were computed using default parameter for alignment of number of ligands matched, volume overlap, relative conformational energy and activity. On basis of survival inactive pharmacophoric model was selected. Among generated models, HHRR_1 with 4 component PLS factor was characterized as best model consist of 34 training set molecules and 17 test set molecules (Table 1). This model is further cross validated by generating two different models of hypothesis as shown in table 2.

 

3D-QSAR Study:

Atom based three-dimensional quantitative structure activity relationship (3D-QSAR) models were generated by dividing data set into training set and test set randomly. These QSAR models are more useful in explaining the structure-activity relationship. The generated pharmacophoric feature HHRR_1 was further used for 3D-QSAR model. We generated an atom based 3D-QSAR model with maximum four component PLS factor. Increase in the number of PLS factors did not improve the model statistics and model predictive ability. We visualized the molecules with generated 3D-QSAR model to correlate their relation between structure and activity.

 

Table 1:  Results of 3DQSAR modelling for training and test molecules

S. No.

Statistical Parameter

Model

1.

No. of training set molecules

34

2.

No. of test set molecules

17

3.

No. of PLS factor

4

4.

SD

0.7137

5.

R2

0.806772

6.

F-value

20.9

7.

Q2

0.7154

8.

Pearson-r

0.8818

9.

RMSE

0.78

 

Validation of 3D-QSAR model:

In the present study, the hypothesis HHRR_1 was characterized as the best hypothesis with a survival score of 5.0328 and the best cross validated R2 0.806772, Q2 0.7154. The generated model produced good Pearson-r 0.8818, RMSE 0.78, F value of 20.9 with PLS factor 4 (Table 1). The generated pharmacophoric model was then validated for its accuracy in prediction of training set molecules activity. The highest correlation of predicted activity of training set molecules was observed with R2 0.806772 which is further validated by cross validation coefficient Q2 0.7154 which implies more confident in the model.


 

Table 2: Summary of atom-based 3D-QSAR results

S. No.

Hypothesis

Survival Score

Survival Inactive

SD

R2

F

P

RMSE

Q2

Pearson-r

1

HHRR_1

5.0328

1.6811

0.7137

0.8067

20.9

0.00027

0.78

0.7154

0.8818

2

HHRR_2

4.9449

1.2969

0.7169

0.8094

19.1

0.000576

0.81

0.7154

0.8681

3

HHRR_3

4.8318

1.1970

0.7694

0.7812

19.6

0.000235

0.70

0.7632

0.9806

 


 

Fig. 1 Alignment of active molecules

 

 

Fig. 2 Alignment of inactive compounds

 

Protein preparation, Ligand Preparation and docking studies:

The crystal structure of protein with PDB ID 2BMA was retrieved from protein data bank site with 2.7 Aº resolution. This protein was optimised by using protein preparation wizard module of Schrodinger Suite 16. Charges and hydrogen were added during this optimization using OPLS force field using restrained minimization. Whereas to refine the side chain and build breaks in between structure Prime module was used. Ligands were prepared by using Ligprep module where they were desalted and tautomers were generated for them. Epik module was used to generate multiple ligand ionisation states and tautomeric sates at pH7.0±0. The strength of association of ligand with receptor is depend on the ligand receptor interaction and docked poses of the ligand with receptor. The docking studies was performed for entire data of set molecules in flexible mode using Glide module of Schrodinger.

 

RESULTS AND DISCUSSION:

The best 4-point HHRR_1 common pharmacophore model was generated based on 34 active compounds from the data set. The statistical data for best generated model including 34 training set and 17 test set showed survival score 5.0328, best R2 0.806772, Q2 0.7154, Pearson-r 0.8818, RMSE 0.78, SD 0.7137 and F value of 20.9. the fitness score was also checked for all ligands with pharmacophore model. The alignment of all the ligands was done by using Flexible ligand alignment tool. These aligned ligands were further used for build-up of 3D-QSAR as shown in Fig.1 where the active ligands mapped with same pharmacophore, the inactives are shown in Fig2. It is an important information from result that all generated hypotheses encompass hydrophobicity and aromatic ring suggesting an important role in the antimalarial activity.

 

Analysis of H-bond donor effect:

3D-QSAR analysis provides a better visualization of ligand receptor interactions. The result helps to visualize favoured region as blue cubes and unflavoured region as red cubes. The blue cubes around pharmacophoric region indicated the preference of hydrogen bond donor groups at those position. Pharmacophoric Model HHRR_1gives 4 regions hypothesis model. The best active molecules fall on these frames. The regions containing rings can be visualized in the figure 3a. The analysis of pharmacophoric regions supports that nitrogen at second position of the Pyrazoline skeleton should have hydrogen donor group. In PYRA_M8 (logIC50= 10.39), PYRA_M10 (logIC50= 10.39), PYRA_M1 (logIC50= 10.30) have nitrogen at second position which acts as hydrogen bond donor (R9 pharmacophoric feature). The presence of unfavourable region for hydrogen bond donor is observed at third position which is supported by the presence of methyl group in PYRA_M8, PYRA_M10. Another observation in this regard is presence of hydrophobic pharmacophoric region in the second pyrazoline ring which is very near to the place of substitution of methyl group. Hydrogen bond donor regions (both favourable and unfavourable) lie very close to the pharmacophoric feature of the first pyrazoline ring. This indicates their impact on the activity and it complies with the requirement of the pharmacophoric model. There is an attachment of phenyl ring through oxygen at position five which is near to the region unfavourable for hydrogen bond. The presence of unfavourable hydrogen bond donor substitution region around pyrazoline ring i.e. R9 pharmacophoric feature (fig.3b) is supported by remarkable decrease in activity of compounds PYRA_M33 (logIC50=7.119) and PYRA_M41 (logIC50=7.14) by three fold.

 

Fig. 3: Pictorial representation of the cubes generated using the 3D-QSAR model and visualized in the context of hydrogen bond donor effect:

 

Analysis of hydrophobic effect:

The presence of blue cubes as shown in fig 4a for compounds PYRA_M8 (logIC50= 10.39), PYRA_M10 (logIC50= 10.39), PYRA_M1 (logIC50= 10.30) around pyrazoline ring substitution (R9 pharmacophoric feature) supported by presence of methyl group at third position which indicated positive effect on activity by hydrophobicity. Secondly other pyrazoline ring at fourteenth position supports hydrophobicity which lies near to hydrophobic region according to pharmacophoric hypothesis. Hence these have a positive effect on activity with all these favoured evidences.  Adding to this assumption the substitution near to oxygen present on fifth position also showing increase in hydrophobic effect.  Hydrophobic groups are not favourable at fifth position of pyrazoline ring for compounds PYRA_M33 (logIC50=7.119) and aromatic ring in PYRA_M 41 (logIC50=7.14) (R12 pharmacophoric feature) as shown in Fig. 4b which is an evidence by the fact that these compounds exhibited less affinity.

 

Fig. 4 Pictorial representation of the cubes generated using the 3D-QSAR model and visualized in the context of hydrophobic effect.

 

Analysis of electron withdrawing effect:

The presence of blue cubes in given fig.5a indicated that compounds PYRA_M8, PYRA_M10 and PYRA_M1 showing favourable region near to R9 pharmacophoric feature because of presence of nitrogen in pyrazoline ring at first position and oxygen at fifth position. This evidence is also supported by the fact that these compounds i.e. PYRA_M8 (logIC50= 10.39), PYRA_M10 (logIC50= 10.39), PYRA_M1 (logIC50= 10.30) showed good affinity. The red cubes showing unfavourable region near to R9 pharmacophoric feature in compounds PYRA_M33 and PYRA_M 41 which is supported by less affinity of these

 compounds (Fig. 5b)

 

Fig. 5: Pictorial representation of the cubes generated using the 3D-QSAR model and visualized in the context of electron withdrawing effect.

 

Docking studies:

Molecular docking studies was performed by using Schrodinger’s Maestro program 11. Ligand candidates with the best conformational and energetic results were selected.  Molecular interaction studies between protein and ligands were performed. The docking studies were performed by using Schrödinger Glide. The downloaded protein structure was refined using Protein preparation wizard. The structures were drawn using 2D sketcher of Schrödinger maestro 11. The ligands were optimised for the docking using LigPrep where the structures were converted to 3D and energy optimised using OPLS3. The fully optimized protein structure was used for the receptor grid generation using Glide software. The prepared ligands were docked to protein grid using extra precession (XP) mode. The protein-ligand complex was further analysed and visualized using Maestro.

 

In the docking studies, pyrazoline derivatives have shown their interaction with pocket amino acids present in protein’s cavity site. The amino acid residues involved interactions with ligands are HIS95, SER97, GLN323, ARG93, ALA321, ALA346, ILE166, ILE102 and PRO96. In general, R9 pharmacophoric feature of compounds showing interaction with SER97, HIS95 and ILE102 amino acids of protein. The best compound that is PYRA _M8 also showed pi-pi stacking and pi-cation interactions with HIS95 and ARG93 respectively. The hydrophobic interacting amino acids involved in ligand protein interaction are ALA321, ALA 346, PRO96, ILE102 and ILE166. Fig.6 (a) shows docking pose and (b) shows ligand interaction diagram of compound PYRA_M8 within the active site of protein. Here nitrogen of pyrazoline ring having R9 pharmacophoric feature accepts a hydrogen bond from SER97. Other pyrazoline ring at fourteenth position also accepts a hydrogen bond from backbone of GLN323 amino acid of protein binding cavity. Whereas pi-cation interaction was formed in between pyrazoline ring of R9 pharmacophoric feature and ARG93. The amino acid residue from backbone of protein binding cavity involved in pi-pi stacking is HIS95. These interactions made compound PYRA_M8 more stable hence more active with good affinty.

 

Fig. 6 Compound PYRA_M8

 

CONCLUSION:

In the present work, an In silico approach was applied to study structure of compounds and their interaction within the binding cavity amino acid residues of selected protein 2BMA. First, we developed a four-point 3D-QSAR model using 80 reported pyrazoline derivatives. This four-point hypothesis consist of two aromatic ring features and two hydrophobic features which provide an information for structural requirement of molecules for further designing of compounds with selective antimalarial activity. Study of correlation between structural features and docking results gave the reason for high activity of compound PYRA_M8 as compare to less active compound PYRA_M33. Whereas developed 3D-QSAR model provides possible structural modifications which will be of interest in designing of more potent and selective pyrazoline derivatives as antimalarial agents.

 

ACKNOWLEDGEMENTS:

We are thankful to Dr Prasahantha from Scientific Bio minds to help us with in silico studies.

 

ABBREVIATIONS:

3D-QSAR, 3-dimensional quantitative structure-activity relationship; PLS, partial least square; RMSD, root mean square deviation; SD, standard deviation; R2, correlation coefficient; Q2, correlation coefficient for test set; XP, extra precision

 

REFERENCES:

1.      Leven M, Held J, Duffy S, Tschan S, Sax S, Kamber J, Frank W, Kuna K, Geffken D, Siethoff, Barth S, Avery VM, Wittin S, Mordmuller B, KurzT. J. Med. Chem. 57, 2014, 7971-7976

2.       Biamonte MA, Wanner J, Roch KGL. Bioorg. Med. Chem. Lett. 23, 2013, 2829-2843.

3.      WHO. 2016, The World Malaria Report from WHO, World Health Organization.

4.      Koca I, Ozgur A, Coskun KA, Tutar Y. Biorg. Med. Chem. 21, 2013, 3859-3865.

5.      Zhu SL, Wu Y, Liu CJ, Wei CY, Tao JC, Liu HM, Eur. J. Med. Chem. 65, 2013, 70-82.

6.      Bekhit A, Hymete A, Bekhit AEDA, A. Damtew, H.Y. Aboul- Enein, Mini Rev. Med. Chem. 10, 2010, 1014-1033.

7.      Nagarapu L, Mateti J, Gaikwad HK, Bantu R, SheebRani M, Subhashini PN, Biorg. Med. Chem. Lett. 21, 2011, 4138-4140.

8.      El-sabbagh OI, Baraka MM, Ibrahim SM, Pannecouque C, Andrei G, Snoeck R, Balzarini J, Rashad A, Eur. J.Med. Chem. 44, 2009, 3746-3753.

9.      Wu L, Song B, Bhadury PS, Yang S, Hu D, Jin L, J. Heterocycl. Chem. 48, 2011, 389-396.

10.   Ragavan RV, Vijayakumar V, Kumari NS, Eur. J. Med. Chem. 45, 2010, 1173-1180.

11.   Vijesh A, Isloor AM, Shetty P, Sundershan S, Fun HK. Eur. J. Med. Chem. 62, 2013, 410-415.

12.   Cunico W, Cechinel CA, Bonacorso HG, Martins MAP, Zanatta N, de Souza MVN, Freitas IO, Soares RPP, Krettli AU. Biorg. Med. Chem. Lett. 16, 2006, 649-653.

13.   Bekhit AA, Hymete A, Damtew A, Bekhit AEDA, Mohamed AMI. J. Enzyme Inhib. Med. Chem.27, 2012, 69-77.

14.   Dias LRS, Salvador RRS, Pharmaceuticals 5, 2012, 317-324.

15.   dos Santos MS, Oliveira ML, Bernardino AM, de Leo RM, Amaral VF, de Carvalho FT, Leon LL, Canto-Cavalheiro MM. Biorg. Med. Chem. Lett. 21, 2011, 7451-7454.

16.   Marra RK, Bernardino AM, Proux TA, Charret KS, Lira MLF, Castro HC, Souza AM, Oliveira CD, Borges JC, Rodrigues CR, Canto-Cavalheiro MM, Leon LL, Amaral VF. Molecules, 17, 2012, 12961-12973.

17.   Dardari Z, Lemrani M, Sebban A, Bahloul A, Hassar M, Kitane S, Berrada M, Boudouma M. Arch. Pharm. Chem. Life Sci. 339, 2006, 291-298.

18.   Garcia-Bustos JF, Francisco-Javier Gamo, Curr. Pharm. Des., 2015, 19, 270.

19.   Chaturvedi D, Goswami A, Saikia PP, Barua NC Rao PG, Chem. Soc. Rev., 2010, 39, 435.

20.   Banek K, Lalani M, Staedke SG, Malar CD. J., 2014, 13, 7.

21.   Bannister LH, Proc. Natl. Acad. Sci. U. S. A., 2001, 98, 383.

22.   Schlitzer M, Chem Med Chem, 2007, 2, 944.

23.   Mckerrow JH, E. Sun E, Rosenthal PJ, Bouvier J, Annu. Rev. Microbiol., 1993, 47, 821.

24.   Liu J, Istvan ES, Gluzman IY, J. Gross and D. E. Goldberg, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 8840.

25.   Karad SC, Purohit VB, Raval DK, European Journal of Medicinal Chemistry, 84, 2014, 51-58.

26.   Bekhit AA, Hassan AMM, Razik HAAE, El-Miligy MMM, El-Agroudy EJ, Bekhit AEDA European Journal of Medicinal Chemistry, 94, 2015, 30-44.

27.   Bekhit AA, Hymete A, Sahille HS, Bekhit AEDA, Arch. Pharm. Chem. Life Sci. 2012, 345, 147–154.

28.   Pandey AK, Sharma S, Pandey M, Alam MM, Shaquiquzzaman M, Akhter M, European Journal of Medicinal Chemistry, 123, 2016, 476-486.

 

 

 

Received on 28.06.2021         Modified on 23.12.2021

Accepted on 30.03.2022   ©Asian Pharma Press All Right Reserved

Asian J. Pharm. Res. 2022; 12(2):119-124.

DOI: 10.52711/2231-5691.2022.00018