INTRODUCTION:

Alzheimer disease (AD), a progressive degenerative disorder of the brain resulting in the loss of higher cognitive function is the most common form of dementia highly prevalent in old age^1–3. The cognitive impairments resulting from AD include agnosia, apraxia and dysphasia^1,4,5.

According to the World Health Organisation (WHO) dementia fact sheet, > 55 million people are currently diagnosed with dementia worldwide and there are estimated 10 million new cases every year⁶. On top of that, dementia is currently the seventh leading cause of death among all diseases and is amidst one of the major causes of disability and dependency among the elderly population globally⁶. In Malaysia, the prevalence of dementia is estimated at 0.454% in 2050⁷.

To date, there is still no effective treatment for AD capable of curing, halting, or slowing down disease development. Current treatment options for AD are rather palliative, addressing the symptoms to provide some temporary relief and enhance the quality of life for patients, but not changing the course of the illness or the rate of decline. Medicines currently prescribed for AD are mainly inhibitors of acetylcholinesterase (AChE) and antagonist of the N-methyl-D-aspartate (NMDA) receptor^1,3,8–11. AChE inhibitors have shown efficacy by temporarily slowing the disease progression, these agents act by inhibiting the enzyme AChE². The NMDA receptor antagonist on the other hand, results in Ca²⁺ influx to activate signal transduction that triggers gene transcription, essential in synaptic neurotransmission, plasticity, and memory formation².

In recent years, the focus on research has been based on the amyloidogenic hypothesis whereby several agents have been reported to successfully inhibit Aβ aggregation. However, this approach has proven to remain only modestly effective as these agents are all characteristically large molecular weight peptides which have limited access across the blood-brain-barrier (BBB)⁸. That said, although this approach may work in principle, the development of a useful drug from this approach may be severely limited due to the issues with regards to access of the drug to the brain. While much fewer drug trials have been focused on the tau hypothesis⁸, with the difficulties encountered with anti-Aβ strategies, interest in tau-related therapeutics has been growing steadily in the recent years^9,10. Consequently, suggest that tau related therapies may be a more viable approach to AD treatment.

To increase the chances of successfully identifying potential disease modifying treatment for AD, the drug repurposing approach have been explored. Drug repurposing offers a potentially time-efficient and cost-effective approach in identifying potential candidates for treatment of AD¹¹. Importantly, these potential candidates have established safety profiles¹² to speed up the entire process of rationale drug discovery phase in complement to the traditional industry-based drug development programmes. There have been extensive successful cases on drug repurposing for several disease areas such as cancer, irritable bowel syndrome, obesity, psychosis, smoking cessation^12–14. Additionally, the drug repurposing method used in AD accounted for ~30% of all AD trials worldwide¹². However, there is still no study conducted on repurposed drugs function as HP tau aggregation inhibitor which is a safer and more viable treatment approach for AD.

Studies that utilised the drug repurposing approach have been observed to be coupled with Structural Bioinformatics methods involves molecular docking^15–22 and molecular dynamics simulation²³. Molecular docking simulation is used to dock potential ligands on target protein thus determining its binding affinities. MD simulations predict the movement of every atom in a system over time based on Newtonian physics governing the interatomic interactions^24,25.

In this work, the unavailability of high resolution of the HP-tau protein structure motivated us to perform multiple template homology modelling to predict the three-dimensional protein structure. The binding site of HP-tau was then predicted using various methods to ensure accuracy, docking at the correct binding site helps reflect the true interaction between the drug and the protein. Screening was then conducted on the approved small molecule drugs from DrugBank, followed by molecular docking to investigate the binding affinities of the shortlisted repurposed drugs in inhibiting tau protein. The two most potential repurposed drug candidates were then subjected to MD simulation to ascertain their potentials as repurposed drug candidates on HP tau protein that could inhibit the aggregation of HP tau.

MATERIALS AND METHODS:

Homology Modelling:

The entire three-dimensional (3D) structure of HP-tau is currently not available and only parts of it can be found in Protein Data Bank²⁶. Since the tau protein is an intrinsically disordered protein it would be difficult to predict the model based on its entire 441 residues. The conserved site of the protein is the microtubule binding region as identified in NCBI. Basic Local Alignment Search Tool for protein (BLAST-P)²⁷ was performed against PDB using residue 252 to 367 as input, which was obtained from NCBI protein database, and the Reference Sequence was NP_005901.2. Based on these, homology modelling was performed on the conserved region. Although tau protein is not only specific to AD, but the structure is disease-specific hence only crystal structures from AD were chosen as templates. The templates were structure of paired helical filaments from AD brain (PDB ID: 6HRE) and tau protein bounded to microtubule (PDB ID: 2MZ7). Multiple template homology modelling was performed using MODELLER 9.25²⁸ for 100 runs and the best protein model was evaluated. Another 100 runs of loop modelling were done on parts of the model that has Discrete Optimised Protein Energy (DOPE) score lower than the templates. The best model after loop refinement was evaluated again and results of the protein validation were compared.

Protein Model Validation:

Protein validation was done using ERRAT²⁹, Verify3D³⁰, PROVE³¹ and PROCHECK³² available on Structural Analysis and Verification Server (SAVES 6.0) (https://saves.mbi.ucla.edu/). Lastly, PROCHECK analyzes the structure geometry of the protein using Ramachandran plot and also calculates the G-factor of the model.

Binding Site Prediction and BBB Penetration Prediction:

The final protein structure was submitted to 5 different servers for binding site prediction, which were DEPTH³³, FTSite³⁴, COACH-D³⁵, ProBis-CHARMMing³⁶ and ConSurf³⁷. The results from each server were considered and the consensus result was used for docking. DrugBank database ver 5.1.7³⁸ of small molecule drugs was downloaded which contained 2,636 drugs. After removing experimental, investigational, illicit drugs and drugs without SMILES or InChl format, the rest of the drugs were screened in batch using LAZAR³⁹ to predict their BBB penetration ability. LAZAR (lazy structure activity relationship) creates local quantitative structure activity relationship (QSAR) for the query compounds and make prediction using data mining algorithms on the dataset of known compounds. The drugs that were predicted to pass the BBB were converted to pdb format using OpenBabel⁴⁰.

Molecular Docking Simulation:

Each shortlisted drug was subjected to 50 runs using AutoDock 4.2.6⁴¹, including 2 controls (curcumin and PE859). These two compounds were found to be effective in reducing aggregation of hyperphosphorylated tau protein in vivo and PE859 is a modified version of curcumin derivative⁴². The docking parameters were set as default and Lamarckian Genetic Algorithm was used. Each drug was set to dock on the binding site for 50 runs. The grid box was 60 × 50 × 55 Å to completely cover the binding site. Drugs that have their lowest binding energy similar to the controls were deemed to be good potential inhibitors and were subject to protein-ligand interaction analysis using LigPlot+⁴³. Best ranked docked compounds then proceed to subsequent MD simulation studies.

Molecular Dynamics Simulation:

A total of 20 ns all-atom MD simulation was performed on HP tau protein and the best-ranked docked compounds (Teniposide bound to HP tau protein, Testosterone Enanthate bound to HP tau protein), each of these systems are referred to as tau, tentau, and testau, respectively throughout this study. All file preparation, minimisation stages and MD productions were performed using AMBER 21⁴⁴. All simulations were performed using the graphics processing unit (GPU) server owned by Swinburne University of Technology, Australia.

General AMBER force field was used to describe the ligands in tentau and testau while the AMBER force field 14SB was used to simulate the MD of the protein in tau, tentau and testau. Next, to ensure neutrality of the systems, 10 negatively charged chlorine ions were added to the negatively charged tau and testau while 11 Cl- were added to tentau. The entire system for all structures were then solvated into a truncated octahedron box of TIP3P water box. The solvated systems underwent two minimisation stages for the purpose of removing bad steric clashes and contacts that exist due to solvation. Three stages - heating stage, NVT equilibration stage and finally NPT production stage. Periodic boundary simulations based on Particle Mesh Ewald with cut-off of 10 Å were used to correct short- and long-range interaction. SHAKE algorithm was also turned on throughout the simulations to constrain fast motion bond that involve hydrogen atoms. Langevin dynamics with the collision frequency gamma_ln set to 1.0 was used to maintain constant temperature. A time step of 2 ps was used in all three stages to allow integration of force equation. The entire system was gently annealed from 0 K to 300 K and the production stage was conducted for 20 ns under NPT at 300 K. The trajectories from the MD simulations were analysed using the CPPTRAJ module⁴⁵ embedded in AmberTools21 and the structure was visually examined and illustrated using Visual Molecular Dynamics (VMD)⁴⁶. Xmgrace⁴⁷ plotting tool was used to generate all the graphs in this study. Analysis of MD trajectories was focused on the production stage (3 ns - 20 ns) while thermodynamics properties were monitored throughout the simulation.

RESULTS AND DISCUSSION:

Homology Modelling:

There are a total of 441 residues consisting of two acidic regions (A1, A2) follow by a proline-rich region and then follow by 4 repeat regions (R1, R2, R3, R4) in the tau protein. The 4 repeats, from residue 244 to residue 367, are identified as the functional region of the protein. There is currently no available protein structure from the database for the first repeat region for AD, and therefore the model is only consisting of the second to fourth repeat region (residue 267 to 380) where 2MZ7 provided structure for residue 267 to 312 and 6HRE provided structure from residue 304 to 380. These two protein structures were chosen as templates for homology modelling, one covering the second and third repeat region of the conserved domain (PDB ID: 2MZ7) shared 100% identity with 1e^-25 E-value and the other covering the third and fourth repeat region (PDB ID: 6HRE) sharing 100% identity with E-value of 1e^-94.

Multiple sequence alignment of the tau conserved region (second, third and fourth repeat only), 2MZ7 and 6HRE was done using Clustal Omega algorithm⁴⁸. This alignment was used in MODELLER to build the protein model. A total of 100 models were built using multiple template homology modelling approach. Out of the 100 models built, model number 62 has the lowest Discrete Optimised Protein Energy (DOPE) of -6226.99 and the normalised DOPE score is -3.63. DOPE is a statistical potential used to assess the energy of the protein model generated, based on that native structure has the lowest free energy of all states under the native conditions⁴⁹. Native conditions means that the protein is properly folded which makes it operative and functional. Hence, the structure that has the lowest DOPE score would be the best probable native structure. The frequency distribution of the DOPE scores ranges from -3.6 to 1.6 which indicates that the quality of the 100 models have high variability. Model 62 is 3 standard deviations away from the mean score (normalised DOPE score = 0) and seems to be an outlier from the distribution. The DOPE score per residue shows that model 62 has higher score than the template at residue 18 to 30, this shows the protein structure at this region has a lower quality than the template. Thus, loop refinement was done on these 13 residues with another 100 runs to adjust the structure and improve the model accuracy. The lowest DOPE score achieved after loop refinement was -762.879 and normalised DOPE score of -0.86. This means the lowest DOPE score is around the mean score (normalised DOPE score = 0) and since the category it is in (-0.68, -0.11) has the highest frequency, this model was not an outlier. Although, there is also a huge range of the normalised DOPE score, from -0.86 to 5.14, the range was like the last protein modelling range, which reflects some consistency of the tool. Tools that could function consistently gives more reliable results.

Protein Model Validation:

The final model (after loop refinement) was subjected to structural validation using a few methods as mentioned in Methods and Material Section. Ramachandran plot statistics obtained from PROCHECK showed significant improvement with no residues on disallowed region and increased in percentage of residues on favoured region. This is in contrast with regions (residues 25 to 37) with higher DOPE score per residue after loop refinement. The difference in G-factor for both before and after the loop refinement was only 0.05, both values are above -0.5 so the structures are not unusual³². Table 1 shows the comparison of protein before and after loop refinement.

Table 1: Ramachandran plot statistics of final model before and after loop refinement. It shows that there is an improvement in the protein structure after loop refinement.

Ramachandran Plot Statistics	Before loop refinement	After loop refinement
Residues in favoured region	80.2%	86.8%
Residues in additional allowed region	9.9%	6.6%
Residues in generously allowed region	7.3%	6.6%
Residues in disallowed region	2.2%	0.0%
G-factor	-0.25	-0.3

Table 2 shows other protein structure validation servers results. ERRAT calculates the percentage of protein that falls below the 95% rejection limit and around 95% ERRAT score indicates high quality²⁹. Verify3D computes the percentage of 3D/1D compatibility of the protein, and a satisfactory model is expected to score around 80%⁵⁰. For these two metrics the higher the scores the better the protein structure would be. The results from ERRAT did not change much from template to protein modelled. For 2MZ7, ERRAT and Verify3D could not get a score for it due to it only contains 45 residues. Verify3D could not get a result from 6HRE as well, but by comparing between the models built, the model before loop refinement was slightly better. PROVE calculates the percentage of buried outliers’ protein atoms, percentages between 1% to 5% is still acceptable but anything higher than 5% can be problematic³¹. For PROVE, the lesser the score, the better the protein structure is. The scores given by PROVE server shows 0.01% increase after loop refinement but a huge increase from the templates. This could be due to the size of the protein as well, each chain of 6HRE contains 76 residues. The protein structure after loop refinement is acceptable and proceed to be used in binding site prediction.

Table 2: Protein structure validation scores of final model before and after loop refinement. Comparing the scores of the protein model before and after loop refinement, the model before loop refinement is slightly better.

Protein Structure Validation Servers	Final Model		Template from 2MZ7	Template from 6HRE
Protein Structure Validation Servers	Before loop refinement	After loop refinement	Template from 2MZ7	Template from 6HRE
ERRAT	59.6%	58.3%	-	58.4%
Verify3D	24.6%	22.8%	-	-
PROVE	13.8%	13.9%	10.6%	8.0%

Binding Site and BBB Penetration Prediction:

Figures 1a-1c show the binding sites predicted by each server which are highlighted in blue. FTSite uses probes to detect for interactions at different site of the protein³⁴ and ProBis-CHARMMing look for similarity between query protein surface and known protein surface that are binding sites³⁶. FTSite and ProBis-CHARMMing predicted 3 binding sites each but those sites are very close to each other while COACH-D predicted 4 binding sites, but those sites are not clustered together. COACH-D uses five different algorithms to predict the binding sites³⁵. ConSurf shows residues that are evolutionary conserved, and the residues predicted on the protein built were not near to each other. The consensus binding site is shown in Figure 1c (right) where it is located at the second and third repeat region and would be used as binding site for molecular docking. There are 1185 of drugs that are Food and Drug Administration approved with SMILES format or InChI format, Figure 2 shows the results of LAZAR screening where only 454 drugs have enough information measurable by LAZAR. Out of these 454 drugs, only forty-two drugs were marked as being able to penetrate the BBB with high confidence, where the probability of penetrating BBB is higher than the probability of not penetrating BBB. These forty-two drugs are used in molecular docking.

Molecular Docking:

Of those forty-two compounds obtained in LAZAR, only ten compounds with lowest binding energy (BE) that is similar to the controls, PE859 and curcumin. Table 3 lists the lowest BE for ten compounds against control dock compounds. The number of residues in the protein that interacted with each drug were compared with the controls, curcumin and PE859. The filtered ten drugs can be grouped into four categories, where five of them were sex hormones variants (Methyltestosterone, Deoxycorticosterone acetate, Testosterone cypionate, Testosterone enanthate, Progesterone), two were antipsychotic drugs (Paliperidone and Nitrazepam), two are antibiotics (Ampicillin and Benzylpenicillin) and one was an anti-tumour drug (Teniposide). The protein-ligand interactions that could be found in the docking studies include forming hydrogen bond (HB) and hydrophobic contacts with the ligand.

Root Mean Square Deviation (RMSD) demonstrates how a conformation deviates from a reference structure as a function of time⁵³ and is the most used quantitative measure of similarity between two atomic coordinates⁵⁴. The position of the entire configuration was compared to the average configuration of all trajectories from 3 ns onwards. The RMSD calculation was done based on the carbon-alpha backbone (Cα), carbon (C) and nitrogen (N) atoms of the entire conformation. Figures 3A-B shows the comparison of RMSD on the backbone atoms for entire simulation system and binding sites versus time for tau, tentau, and testau as a function of time from 3 ns to 20 ns. Based on the average RMSD, starting frames of the simulation for all 3 systems showed much higher RMSD values, while after approximately 7 ns, stabilisation of the protein was observed, however, fluctuations were visible over the course of the simulation. The RMSD analysis was also performed on the binding site residues whereby the RMSD calculation was done based on the Cα, C and N atoms of the overlapping residues with 10 Å distance from the ligands based on the average trajectory. It can be observed that the RMSD trend for tentau was relatively constant with minor to no fluctuation from 3 ns to 18 ns. Overall, the RMSD values for tentau was mostly below the figures of tau while testau had significantly higher RMSD values throughout the simulation.

The Root Mean Square Fluctuation (RMSF) measures the displacement of a particular atom, or group of atoms⁵⁵, and the flexibility of the different regions of a protein⁵⁶, relative to the average deviation over time from a reference position. The RMSF analysis of the MD simulation was performed based on the Cα, C and N atoms for each residue from 3 ns to 20 ns trajectories of the tau, tentau and testau system. The tau system acted as the standard and was used for comparison purposes as this system did not contain any ligand. The RMSF graph (Figure 4A) demonstrated per-residue fluctuations for tentau behaves in parallel or similarly to tau for all residues throughout the simulation. The low RMSF values for tentau suggest that the residues are more rigid due to the presence of strong interactions with the ligand. On the contrary, the RMSF values for testau deviates extensively from tau, especially from residue number 70 to 95, suggesting that these residues are extremely flexible with high fluctuations and contributed the most to the molecular motions during MD simulation of testau. Thus, indicated that these residues of testau were possibly distant from the vicinity of the binding site. The conformational flexibilities of the drug-receptor complexes can be further examined from evaluating the RMSF of the residues within the area of the binding site. That said, the overlapping residues that were 10 Å distance away from the ligand based on the average trajectory of tentau and testau can be considered as among the key residues to be evaluated and compared based on the RMSF values. The overlapping binding site residues included SER19, ASN20, VAL21, GLN22 and LYS51.

Figure 3: (A) Comparison of RMSD versus time for tau, tentau and testau entire system. (B) Comparison of RMSD versus time for tau, tentau, and testau overlapping binding site residue.

Figure 4: Comparison of RMSF per-residue (A), RoG (B) for tau, tentau, and testau, respectively.

The Radius of Gyration (RoG) acts as an indicator on the compactness and the size of protein molecules⁵⁷ as the compactness of a protein has a direct relationship with the rate of folding⁵⁸. An increase in RoG values indicates a decrease in structure compactness, thereby proposing less stability and increased flexibility of the system. The RoG analysis of the MD simulation was performed based on the Cα, C and N atoms of the overlapping residues with 10Å distance from the ligands based on the average trajectory. The RoG was measured to investigate the binding stability of Teniposide and Testosterone Enanthate to the HP tau protein. Figure 4B illustrates the RoG analysis of the backbone atoms of the overlapping residues for tentau and testau in comparison with tau as a function of time from 3 ns to 20 ns. Similar to the RMSD and RMSF analysis, the tau system acted as the standard and was used for comparison purposes as this system did not contain any ligand. By comparing the graphs for both tentau and testau with tau, it is clearly evident that the binding of both ligands resulted in the increase in protein structure compactness on the HP tau protein as the RoG values for both systems were overall lower than of tau.

CONCLUSION:

AD is devastating as the disease causes an irreversible progressive brain degeneration and is considered as among the most common form of dementia highly prevalent in old age. The current study successfully predicted the conserved domain of the hyperphosphorylated tau protein using homology modelling. Screening was done on the approved small molecular drugs from DrugBank and 42 drugs were predicted to pass blood brain barrier. The shortlisted Teniposide and Testosterone Enanthate through drug repurposing, and its binding affinities were examined by molecular docking and MD simulation. Docking results found that Teniposide had the lowest binding energy at -8.15 kcal/mol while Testosterone Enanthate was -8.20 kcal/mol. MD simulation provided quantitative and qualitative detailed structural analysis for both drugs. MD simulation helps to fill in the gaps by allowing flexibility and movement relating to the stability of the protein-ligand complex interaction. The findings from this study can be used as stepping stone to investigate the binding interactions of potential anti-aggregation agents with HP tau protein; from repurposed drugs to inhibit the aggregation of HP tau. The significance of this study can be further emphasised on the grounds that there are currently no drugs available in the market capable of curing, halting, or slowing down AD development as the current treatment strategies are centred towards palliative care and do not change the course of the illness. The limitation of this work came from the relatively short MD simulation time and limited analysis in fine tuning the docking results. Thus, the scope of additional future work should be focused on conducting a longer simulation to improve reliability. Additional physico-chemico analysis would definitely be fundamental to assist in this endeavour to explore potential compounds from repurposed drugs in AD.

REFERENCES:

1. Martone RL, Piotrowski NA. Alzheimer’s Disease, Magill’s Medical Guide (Online Edition). Salem Press; 2022.

2. Breijyeh Z, Karaman R, Muñoz-Torrero D, Dembinski R. Comprehensive Review on Alzheimer’s Disease: Causes and Treatment. Molecules. 2020; 25(24): 5789. doi:10.3390/MOLECULES25245789

3. Weller J, Budson A. Current understanding of Alzheimer’s disease diagnosis and treatment. F1000Res. 2018; 7. doi:10.12688/F1000RESEARCH.14506.1

4. Sanmugam K. Depression is a Risk Factor for Alzheimer Disease-Review. Res J Pharm Technol. 2015; 8(8): 1056. doi:10.5958/0974-360X.2015.00181.X

5. Jadhav RP, Kengar MD, Narule O V., Koli VW, Kumbhar SB. A Review on Alzheimer’s Disease (AD) and its Herbal Treatment of Alzheimer’s Disease. Asian Journal of Research in Pharmaceutical Science. 2019; 9(2): 112. doi:10.5958/2231-5659.2019.00017.1

6. World Health Organization (WHO). Dementia. Fact sheets. Published September 2, 2021. Accessed June 12, 2022. https://www.who.int/news-room/fact-sheets/detail/dementia

7. Alzheimer’s Disease International (ADI), Dementia Australia. Dementia in the Asia Pacific Region. Access Economics PTY Limited. Published November 1, 2014. Accessed June 12, 2022. https://www.alzint.org/resource/dementia-in-the-asia-pacific-region/

8. Standridge JB. Pharmacotherapeutic approaches to the treatment of Alzheimer’s disease. Clin Ther. 2004; 26(5): 615-630. doi:10.1016/S0149-2918(04)90064-1

9. Vaz M, Silvestre S. Alzheimer’s disease: Recent treatment strategies. Eur J Pharmacol. 2020; 887: 173554. doi:10.1016/J.EJPHAR.2020.173554

10. Folch J, Petrov D, Ettcheto M, et al. Current Research Therapeutic Strategies for Alzheimer’s Disease Treatment. Neural Plast. 2016;2016. doi:10.1155/2016/8501693

11. Corbett A, Williams G, Ballard C. Drug Repositioning: An Opportunity to Develop Novel Treatments for Alzheimer’s Disease. Pharmaceuticals. 2013; 6(10): 1304-1321. doi:10.3390/PH6101304

12. Ihara M, Saito S, Friedland R. Drug Repositioning for Alzheimer’s Disease: Finding Hidden Clues in Old Drugs. J Alzheimers Dis. 2020; 74(4): 1013-1028. doi:10.3233/JAD-200049

13. Ashburn TT, Thor KB. Drug repositioning: identifying and developing new uses for existing drugs. Nat Rev Drug Discov. 2004; 3(8): 673-683. doi:10.1038/NRD1468

14. Senthil Venkatachalam AJADRKVijayakumar. Repurposing Drugs for Management of Alzheimer Disease. . Research J Pharm and Tech. 2019; 12(6): 3078-3088.

15. S. Patil V, A. Patil P. Molecular Docking: A useful approach of Drug Discovery on the Basis of their Structure. Asian Journal of Pharmaceutical Research. Published online September 7, 2023: 191-195. doi:10.52711/2231-5691.2023.00036

16. Patadiya N, Vaghela V. Design, in-silico ADME Study and molecular docking study of novel quinoline-4-on derivatives as Factor Xa Inhibitor as Potential anti-coagulating agents. Asian Journal of Pharmaceutical Research. Published online August 12, 2022: 207-211. doi:10.52711/2231-5691.2022.00034

17. Hemalatha K, Selvin J, Girija K. Synthesis, In silico Molecular Docking Study and Anti-bacterial Evaluation of some Novel 4-Anilino Quinazolines. Asian Journal of Pharmaceutical Research. 2018; 8(3): 125. doi:10.5958/2231-5691.2018.00022.9

18. Bhardwaj S, Dubey S. Qsar and Docking Studies of Some Novel Piperine Analogues as Monoamine Oxidase Inhibitors. Asian Journal of Research in Pharmaceutical Sciences. Published online May 25, 2022: 115-122. doi:10.52711/2231-5659.2022.00019

19. Devgan M. Homology modeling and molecular docking studies of DNA replication licensing factor minichromosome maintenance protein 5 (MCM5). Asian Journal of Pharmacy and Technology. 2015; 5(1): 17. doi:10.5958/2231-5713.2015.00004.5

20. Thorat BR, Nazirkar B, Thorat VB, More K, Jagtap R, Yamgar R. Synthesis, SAR, Molecular Docking and Anti-TB study of 3-Hydroxy-1-Benzofuran-2-Carbohydrazide. Asian Journal of Research in Chemistry. 2016; 9(3): 116. doi:10.5958/0974-4150.2016.00021.3

21. Govindammal M, Prasath M, Sathya B, Selvapandiyan M. Investigation on the Binding Interaction Between Clomipramine and Doxepin with LeuT by Molecular Docking Analysis. Asian Journal of Research in Chemistry. 2017; 10(4): 486. doi:10.5958/0974-4150.2017.00079.7

22. Bagal A, Borkar T, Ghige T, et al. Molecular Docking – Useful Tool in Drug Discovery. Asian Journal of Research in Chemistry. Published online April 9, 2022: 129-132. doi:10.52711/0974-4150.2022.00020

23. Hollingsworth SA, Dror RO. Molecular dynamics simulation for all. Neuron. 2018; 99(6): 1129. doi:10.1016/J.NEURON.2018.08.011

24. Hollingsworth SA, Dror RO. Molecular dynamics simulation for all. Neuron. 2018; 99(6): 1129. doi:10.1016/J.NEURON.2018.08.011

25. Durrant JD, McCammon JA. Molecular dynamics simulations and drug discovery. BMC Biol. 2011; 9. doi:10.1186/1741-7007-9-71

26. Berman HM, Westbrook J, Feng Z, et al. The Protein Data Bank. Nucleic Acids Res. 2000; 28(1): 235-242. doi:10.1093/NAR/28.1.235

27. Altschul SF, Madden TL, Schäffer AA, et al. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 1997; 25(17): 3389-3402. doi:10.1093/NAR/25.17.3389

28. Webb B, Sali A. Comparative Protein Structure Modeling Using MODELLER. Curr Protoc Bioinformatics. 2016; 54(1): 5.6.1-5.6.37. doi:10.1002/CPBI.3

29. Colovos C, Yeates TO. Verification of protein structures: Patterns of nonbonded atomic interactions. Protein Science. 1993; 2(9): 1511-1519. doi:10.1002/PRO.5560020916

30. Lüthy R, Bowie JU, Eisenberg D. Assessment of protein models with three-dimensional profiles. Nature 1992; 356: 6364. 1992;356(6364):83-85. doi:10.1038/356083a0

31. Pontius J, Richelle J, Wodak SJ. Deviations from Standard Atomic Volumes as a Quality Measure for Protein Crystal Structures. J Mol Biol. 1996; 264(1): 121-136. doi:10.1006/JMBI.1996.0628

32. Laskowski RA, Rullmann JAC, MacArthur MW, Kaptein R, Thornton JM. AQUA and PROCHECK-NMR: Programs for checking the quality of protein structures solved by NMR. Journal of Biomolecular NMR 1996 8:4. 1996; 8(4): 477-486. doi:10.1007/BF00228148

33. Tan KP, Nguyen TB, Patel S, Varadarajan R, Madhusudhan MS. Depth: a web server to compute depth, cavity sizes, detect potential small-molecule ligand-binding cavities and predict the pKa of ionizable residues in proteins. Nucleic Acids Res. 2013; 41(W1): W314-W321. doi:10.1093/NAR/GKT503

34. Ngan CH, Hall DR, Zerbe B, Grove LE, Kozakov D, Vajda S. FTSite: high accuracy detection of ligand binding sites on unbound protein structures. Bioinformatics. 2012; 28(2): 286-287. doi:10.1093/BIOINFORMATICS/BTR651

35. Wu Q, Peng Z, Zhang Y, Yang J. COACH-D: improved protein–ligand binding sites prediction with refined ligand-binding poses through molecular docking. Nucleic Acids Res. 2018; 46(W1): W438-W442. doi:10.1093/NAR/GKY439

36. Konc J, Miller BT, Štular T, et al. ProBiS-CHARMMing: Web Interface for Prediction and Optimization of Ligands in Protein Binding Sites. J Chem Inf Model. 2015; 55(11): 2308-2314. doi:10.1021/ACS.JCIM.5B00534/SUPPL_FILE/CI5B00534_SI_001.PDF

37. Ashkenazy H, Abadi S, Martz E, et al. ConSurf 2016: an improved methodology to estimate and visualize evolutionary conservation in macromolecules. Nucleic Acids Res. 2016; 44(W1): W344-W350. doi:10.1093/NAR/GKW408

38. DrugBank. DrugBank Release | DrugBank Online. Published online 2020. https://go.drugbank.com/releases/latest

39. Maunz A, Gütlein M, Rautenberg M, Vorgrimmler D, Gebele D, Helma C. Lazar: A modular predictive toxicology framework. Front Pharmacol. 2013; 4 APR: 38. doi:10.3389/FPHAR.2013.00038/BIBTEX

40. O’Boyle NM, Banck M, James CA, Morley C, Vandermeersch T, Hutchison GR. Open Babel: An Open chemical toolbox. J Cheminform. 2011; 3(10): 1-14. doi:10.1186/1758-2946-3-33/TABLES/2

41. Morris GM, Ruth H, Lindstrom W, et al. AutoDock4 and AutoDockTools4: Automated docking with selective receptor flexibility. J Comput Chem. 2009; 30(16): 2785-2791. doi:10.1002/JCC.21256

42. Okuda M, Hijikuro I, Fujita Y, et al. PE859, a Novel Tau Aggregation Inhibitor, Reduces Aggregated Tau and Prevents Onset and Progression of Neural Dysfunction In Vivo. PLoS One. 2015; 10(2): e0117511. doi:10.1371/JOURNAL.PONE.0117511

43. Laskowski RA, Swindells MB. LigPlot+: Multiple ligand-protein interaction diagrams for drug discovery. J Chem Inf Model. 2011; 51(10): 2778-2786. doi:10.1021/CI200227U/ASSET/IMAGES/MEDIUM/CI-2011-00227U_0006.GIF

44. Case DA, Aktulga HM, Belfon K, et al. Amber 2021: Reference Manual. Published online 2021. https://ambermd.org/contributors.html

45. Roe DR, III TEC. PTRAJ and CPPTRAJ: software for processing and analysis of molecular dynamics trajectory data. J Chem Theory Comput. 2013; 9(7): 3084-3095.

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

47. Turner PJ. XMGRACE, Version 5.1. 19. Published online 2005.

48. Madeira F, Park YM, Lee J, et al. The EMBL-EBI search and sequence analysis tools APIs in 2019. Nucleic Acids Res. 2019; 47(W1):W636-W641. doi:10.1093/NAR/GKZ268

49. Shen M yi, Sali A. Statistical potential for assessment and prediction of protein structures. Protein Science. 2006; 15(11): 2507-2524. doi:10.1110/PS.062416606

50. Khor BY, Tye GJ, Lim TS, Noordin R, Choong YS. The Structure and Dynamics of BmR1 Protein from Brugia malayi: In Silico Approaches. International Journal of Molecular Sciences 2014, Vol 15, Pages 11082-11099. 2014; 15(6): 11082-11099. doi:10.3390/IJMS150611082

51. DrugBank. Paliperidone | DrugBank Online. Published online 2007. https://go.drugbank.com/drugs/DB01267

52. Tsai MC, Chang PT, Yang CH, Liu ME. The delirium related to oral paliperidone in dementia: A case report. J Clin Psychopharmacol. 2016; 36(2): 184-185. doi:10.1097/JCP.0000000000000472

53. Cheng X, Ivanov I. Molecular Dynamics. Published online 2012:243-285. doi:10.1007/978-1-62703-050-2_11

54. Kufareva I, Abagyan R. Methods of protein structure comparison. Methods Mol Biol. 2012; 857: 231. doi:10.1007/978-1-61779-588-6_10

55. Martínez L. Automatic Identification of Mobile and Rigid Substructures in Molecular Dynamics Simulations and Fractional Structural Fluctuation Analysis. PLoS One. 2015; 10(3). doi:10.1371/JOURNAL.PONE.0119264

56. Cheng X, Ivanov I. Molecular Dynamics. Published online 2012:243-285. doi:10.1007/978-1-62703-050-2_11

57. Arnittali M, Rissanou AN, Harmandaris V. Structure Of Biomolecules Through Molecular Dynamics Simulations. Procedia Comput Sci. 2019; 156: 69-78. doi:10.1016/J.PROCS.2019.08.181

58. Sneha P, Priya Doss CG. Molecular Dynamics: New Frontier in Personalized Medicine. Adv Protein Chem Struct Biol. 2016; 102: 181-224. doi:10.1016/BS.APCSB.2015.09.004

Received on 19.06.2024 Revised on 11.09.2024

Accepted on 08.11.2024 Published on 20.01.2025

Available online from January 27, 2025

Research J. Pharmacy and Technology. 2025;18(1):67-75.

DOI: 10.52711/0974-360X.2025.00011

This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. Creative Commons License.

Compounds	Lowest BE (kcal/mol)	Number of overlapped residues interacted with curcumin	Number of overlapped residues interacted with PE859
PE859	-8.84	9	-
Teniposide	-8.17	10	7
Paliperidone	-7.03	10	8
Nitrazepam	-7.71	6	7
Methyltestosterone	-7.00	7	9
Deoxycorticosterone acetate	-7.72	8	7
Testosterone cypionate	-7.95	8	10
Testosterone enanthate	-8.20	8	12
Progesterone	-7.26	8	6
Ampicillin	-7.15	8	6
Benzylpenicillin	-7.15	7	10
Curcumin	-6.71	-	9