Computational prognostic evaluation of Alzheimer’s drugs from FDA-approved database through structural conformational dynamics and drug repositioning approaches

Drug designing is high-priced and time taking process with low success rate. To overcome this obligation, computational drug repositioning technique is being promptly used to predict the possible therapeutic effects of FDA approved drugs against multiple diseases. In this computational study, protein modeling, shape-based screening, molecular docking, pharmacogenomics, and molecular dynamic simulation approaches have been utilized to retrieve the FDA approved drugs against AD. The predicted MADD protein structure was designed by homology modeling and characterized through different computational resources. Donepezil and galantamine were implanted as standard drugs and drugs were screened out based on structural similarities. Furthermore, these drugs were evaluated and based on binding energy (Kcal/mol) profiles against MADD through PyRx tool. Moreover, pharmacogenomics analysis showed good possible associations with AD mediated genes and confirmed through detail literature survey. The best 6 drug (darifenacin, astemizole, tubocurarine, elacridar, sertindole and tariquidar) further docked and analyzed their interaction behavior through hydrogen binding. Finally, MD simulation study were carried out on these drugs and evaluated their stability behavior by generating root mean square deviation and fluctuations (RMSD/F), radius of gyration (Rg) and soluble accessible surface area (SASA) graphs. Taken together, darifenacin, astemizole, tubocurarine, elacridar, sertindole and tariquidar displayed good lead like profile as compared with standard and can be used as possible therapeutic agent in the treatment of AD after in-vitro and in-vivo assessment.


Virtual screening using PyRx
Before conducting our docking experiments, all the screened drugs were sketched in ACD/ChemSketch tool and accessed in mol format.Furthermore, UCSF Chimera 1.10.1 tool was employed for energy minimization of each ligand having default parameters such as steepest descent and conjugate gradient steps 100 with step size 0.02 (Å), and update interval was fixed at 10.In PyRx docking experiment, all screened drugs were docked with death domain of MADD using default procedure 26 .Before, docking binding pocket of target protein was confirmed from Prankweb and literature data.In docking experiments, the grid box dimension values were adjusted as center_X = − 0.8961, Y = − 1.6716 and Z = 0.3732 whereas, size_X = 37.8273, Y = 36.5416Z = 36.5756,respectively, with by default exhaustiveness = 8 value.The grid box size was adjusted on binding pocket residues to allows the ligand to move freely in the search space.Furthermore, the generated docked complexes were keenly analyzed to view their binding conformational poses at active binding site of MADD protein.Moreover, these docked complexes were evaluated based on the lowest binding energy (Kcal/mol) values and binding interaction pattern between ligands and target protein.The graphical depictions of all the docked complexes were accomplished by UCSF Chimera 1.10.1 and Discovery Studio (2.1.0),respectively.
Furthermore, another docking experiment was employed on best screened drugs against MADD protein using AutoDock 4.2 tool 27 .In brief, for receptor protein, the polar hydrogen atoms and Kollman charges were assigned.For ligand, Gasteiger partial charges were designated, and non-polar hydrogen atoms were merged.All the torsion angles for screened drugs were set free to rotate through the docking experiment.A grid map of 80 × 80 × 80 Å was adjusted on the binding pocket of MADD to generate the grid map and to get the best conformational state of docking.A total of 100 number of runs were adjusted using docking experiments.The Lamarckian genetic algorithm (LGA) and empirical free energy function were applied by taking docking parameters default 28 .All the docked complexes were further evaluated on lowest binding energy (kcal/mol) values and hydrogen and hydrophobic interactions analysis using Discovery Studio (2.1.0)and UCSF Chimera 1.10.1.

Designing of pharmacogenomics networks
To design the pharmacogenomics network model for best-selected drugs, Drug Gene Interaction Databases (DGIdb) (https:// www.dgidb.org/) and Drug Signatures Database (DSigDB) (http:// dsigdb.tanlab.org/ DSigD Bv1.0/) were employed to obtain the possible list of different disease-associated genes.Furthermore, a detailed literature survey was performed against all predicted genes to identify its involvement in AD.Moreover, clumps of different diseases associated genes were sorted based on MADD and remaining disease-associated genes were eliminated from the dataset.

Molecular dynamics (MD) simulations
The best screened drugs-complexes having good energy values were selected to understand the residual backbone flexibility of protein structure; MD simulations were carried out by Groningen Machine for Chemicals Simulations (GROMACS 4.5.4package 29 , with GROMOS 96 force field 30 .The protein topology was designed by pdb2gmx command by employing GROMOS 96 force field.Moreover, simulation box with a minimum distance to any wall of 10 Å (1.0 nm) was generated on complex by editconf command.Moreover, box was filled with solvent molecules using gmx solvate command by employing spc216.growater model.The overall system charge was neutralized by adding ions.The steepest descent approach (1000 ps) for protein structure was applied for energy minimization.For energy minimization the nsteps = 50,000 were adjusted with energy step size (emstep) 0.01 value.Particle Mesh Ewald (PME) method was employed for energy calculation and for electrostatic and van der waals interactions; cut-off distance for the short-range VdW (rvdw) was set to 14 Å, whereas neighbor list (rlist) and nstlist values were adjusted as 1.0 and 10, respectively, in em.mdp file 31 .This method permits the use of the Ewald summation at a computational cost comparable with that of a simple truncation method of 10 Å or less, and the linear constraint solver (LINCS) 32 algorithm was used for covalent bond constraints and the time step was set to 0.002 ps.Finally, the molecular dynamics simulation was carried out at 100 ns with nsteps 50,000,000 in md.mdp file.Different structural evaluations such as root mean square deviations and fluctuations (RMSD/RMSF), solvent accessible surface areas (SASA) and radii of gyration (Rg) of back bone residues were analyzed through Xmgrace software (http:// plasma-gate.weizm ann.ac.il/ Grace/).

Results and discussion
The overall computational drug repositioning flowchart showed the multiple steps involved in this study (Fig. 1)..The prior research reports justified that the GRAVY negative and positive values show the hydrophilic and hydrophobic behavior of protein structure 33 .The negative GRAVY value (− 0.517) of MADD protein indicates the hydrophilic behavior (Table 1).

VADAR analysis
VADAR is an online server which predict the protein volume, area, dihedral angle reports and statistical evaluation.The generated MADD model structure showed different architectures related to α-helices, β-sheets, coils and turn, respectively.The overall protein structure analysis showed that MADD consists of 25% α-helices, 19% β-sheets, 55% coils and 20% turns, respectively.Moreover, Ramachandran plots and values depicted that 92.68%  of amino acids were present in favored region with good accuracy of phi (φ) and psi (ψ) angles which showed the good accuracy of our predicted model.

Galaxy refine analysis
Galaxy-refine is an good source to resolve the ambiguities in the modelled structure based on different parameters such as GDT-HA, RMSD, MolProbity, clash score, poor rotamers and Rama favored, respectively 36 .The initial predicted structure of MADD protein depicted MolProbity: 3.809-2.871,clash score: 88.4-47.8,poor rotamers: 5.7-1.4 and Rama favored: 72.4-85.8,respectively whereas, the values have been rectified to improve the modelled MADD structure (Supplementary Table S1).Moreover, the online Ramachandran graph server (https:// www.umass med.edu/ zlab/) generate both Ramachandran graphs which have been mentioned in Fig. 3.
The comparative results showed that the Galaxy refine clearly improved the modelled MADD structure which can be further implied for structure and docking analysis.

FRST energy validation against MADD
The energy validation of predicted MADD structure was analyzed using FRST evaluation tool based on overall and a per-residue energy profile of a protein structure.The FRST energy calculation is composed of rapdf pairwise potential, solvation potential, backbone hydrogen bonds and torsion angle potential and composite score values (http:// prote in.cribi.unipd.it/ frst/).The rapdf potential 37 is a distance-dependent residue-specific all-atom probability discriminatory function.Similarly, solvation potential to likelihood for each of 20 amino acids to adopt a given relative solvent accessibility is used to derive the pseudo-energy 38 .The torsion angle potential (TORS) is derived from the propensities of each amino acid to adopt different combinations of (phi, psi) torsion angles 39 .The total energy is again summed over all individual contributions in a protein.In our generated MADD results, rapdf, solvation, torsion and composite FRST energies (Kcal/mol) which showed the stability of model structure (Table 2).

Quality model energy analysis
Qualitative Model Energy ANalysis (QMEAN) is a composite scoring function which derive both global and local absolute quality estimates based on one single model (https:// swiss model.expasy.org/ qmean/).The generated results showed that MADD protein exhibited − 6.64 value which is comparable with standard values (0,1).The prior data showed that different residues clumps showed lowest score value 0.1 which represents the quality assessment of predicted MADD model.However, most amino acids patches in the predicted model showed higher values > 0.6 depicts accurate and stable behavior of MADD protein.The blue peak showed highest prediction value having best quality results (> 0.8) of predicted model.Moreover, the purple color represents also showed good accuracy of MADD protein with scoring values range from 0.6 to 0.8.However, the orange values showed the de-stable behavior of MADD protein with values less than 0.6.The overall results showed that predicted MADD model exhibited good stable behavior and could be used for further analysis (Fig. 4).

Native conformational behavior evaluation
After prediction the protein models, native conformation behavior is significant parameter is to evaluate the accuracy of protein folds.SPS server (http:// aleph.upf.edu/ spser ver/ index.php/ init.html), is a knowledge-based scoring functions used for correct folding of proteins, stability of mutant proteins, and assess outcomes of docking experiments.The generated results were evaluated based on five different parameters such as PAIR, ES3DC, ECOMB, ELOCAL, and E3DC respectively.The PAIR and ES3DC parameters are focused on amino acid frequencies along distances and their environments such as hydrophobicity of each amino acid, solvent accessibility, and secondary structure, respectively.Figure 5A, B showed that residues fluctuations with respect to z-score values.
In PAIR parameters analysis, residues showed stable frequency throughout the protein structure except amino  showed high fluctuations at the terminal part of MADD protein.However, the average value ranges from 3 to − 3 in the energy folding graph.Our results depicted that MADD protein exhibited hydrophobicity and solvent accessibility properties, respectively.Moreover, both E3DC and ELOCAL parameters are also implicated on MADD protein based on amino acid frequencies along distances of pairs referred by the hydrophobicity of the amino acids and the rest of their environments, whereas ECOMB is combinatorial scores of all three parameters such as ES3DC, ELOCAL and E3DC, respectively.Figure 5C displayed the residual fluctuations graph at different positions in the MADD structure whereas, ELOCAL parameter showed less fluctuations based on Z-score values (Fig. 5D).In combine effects, the ECOMB parameter exhibited stable behavior and good folding of model MADD protein.The fluctuation peaks were range from 0 to 3, however, at various position range raised up to 4 (Fig. 5E).

Binding pocket analysis of MADD protein
Generally, binding pocket is the core position of ligands in the active region of protein 40 .Prankweb is a novel machine learning method for prediction of ligand binding sites inside the protein structure 41    www.nature.com/scientificreports/value 152 as compared to other binding pockets values (66, 49, 28 and 18, respectively) with different amino acids of death domain of MADD protein.Moreover, the probability value of pocket 1 was also better as compared to other probabilities values (Fig. 6A, B).

Shape based screening and retrieval of similar FDA approved drugs
In drug-repositioning approach, shape-based screening, pharmacogenomics and molecular docking simulation are significant parameters to predict the possible drugs from FDA approved medications by considering protein targets 5,42 .Donepezil and galantamine were used as standard drug against AD and used as template to screen FDA approved drug having similar skeleton similarity.In our computational results, Swiss-Similarity results showed 282 drugs were retrieved with donepezil and 351 against galantamine.The scoring values FDA approved drugs based on structural similarity scoring values range from 1.000 to 0.010 and 1.000 to 0.134 for donepezil and galantamine, respectively (Supplementary Tables S2 and S3).The screened drugs were ranked based on similarity scoring values, ranged for 0-1.The 0 value represents dissimilarity between compounds whereas, 1 is used for highly identical compounds in screening approach 5 .In our computational results, drugs depicted good similarity scoring values range from 1 to 0.111.In donepezil screening results, DB07701 and DB3393 were exhibited scoring values 0.999 and 0.837, respectively.Furthermore, in galantamine screening results, it has been observed that DB00318 (0.979), DB01466 (0.977), and DB11490 (0.974) showed scoring values and DB14703 depicted lowest structure similarity result 0.134.The top 100 screened drugs from each standard (donepezil and galantamine) were selected for molecular docking analysis to predict the most suitable candidate having good binding affinity values against MADD protein.

Virtual screening using PyRx
PyRx is a virtual screening software for computational drug discovery that can be used to screen libraries of compounds against potential drug targets.In PyRx, donepezil-MADD docking results, among 100 screened drugs (Table S4), 72 drugs showed higher docking energy values as compared to standard donepezil (Table 3).Similarly, In PyRx, galantamine-MADD docking results, among 100 screened drugs (Table S5), 67 drugs showed higher docking energy values as compared to standard galantamine (Table 4).

Pharmacogenomics analysis
Based on swiss similarity and virtual screening result, the selected drugs further undergo for pharmacogenomics analysis to check their possible interaction with genes encoded proteins and their association with AD.Multiple reports also justify this approach as worthwhile to predict the possible drug having repositioned functionality 5,12 .
All the screened drug from both standard (donepezil and galantamine) exhibited good interaction scoring values with predicted genes which showed good correlation with AD.From donepezil screening, our results showed that 10 from 35 drugs exhibited good interaction scoring values and association with AD.The darifenacin showed good interaction scoring values with CHRM3 (0.85), CHRM2 (0.38), CHRM1 (0.29) and CYP2D6 (0.04) which have correlation with AD.The other drug astemizole showed interaction with multiple genes ( 14) which have association with multiple diseases.The 9 genes such as CYP2J2, HPSE, ABCB1, PPARD, CYP2D6, IDH1, CYP3A4, AR, and TP53 were directly associated with AD.The predicted results showed that astemizole showed highest interaction score 0.90 as compared to other genes.In tubocurarine gene network, 5 genes were interacted and among these four 2 (CHRNA2 and KCNN2) genes were showed interaction with AD.Both CHRNA2 and KCNN2 genes exhibited good interaction scoring values 6.37 and 4.24, respectively.Another important screened drug which exhibited good interaction prediction with AD is aripiprazole through interaction with different genes.The drug-gene association analysis showed that out of 14 genes, eleven (DRD2, HTR1A, ANKK1, SH2B1, HTR2A, CNR1, FAAH, HTR1B, HTR2C, ABCB1, CYP2D6) possessed good interaction scoring values and are directly involves in AD.Fluspirilene-gene interactions analysis also showed their intimation with AD by targeting multiple genes.The predicted results showed that six gene such as DRD2, HTR1E, XBP1, HTR2A, HTR1A, and PPARD were showed their association with AD with different interaction scoring values.In elacridar-genetic complex, two genes ABCG2 (1.54) and ABCB1 (0.27) were observed having showed their interaction with AD.Another important screened drug was sertindole which showed interactions multiple genes involves in AD and other diseases.A total of 17 genes were observed showed interaction with sertindole and from 17, eight genes (DRD2 (0.15), HTR2A (0.14), HTR2C (0.17), HTR1E (0.1), HTR1A (0.03), HTR1B (0.05), CYP2D6 (0.01), CYP3A4 (0.01)) showed association with AD with different interaction scoring values.Tariquidar interacted with ABCB1 and ABCG2 with scoring values 0.66 and 1.03, respectively.Sarizotan is another screened drug shows interactions with HTR1A (0.91) and DRD2 (0.36) having good interaction score values with AD.Similarly, fipexide interacted with CYP2D6 (0.03) and CYP3A4 (0.02) also depicted good correlation with AD (Table 5).

Docking energy analysis against death domain of MADD
The selected drugs-death domain docked complexes were analyzed based on lowest binding energy values (Kcal/mol) and hydrogen/hydrophobic interaction analyses.Results showed that all drugs showed good binding energy values (Kcal/mol) and binds within the active site with appropriate conformational poses (Table 7).Docking energy values is most significant parameter to screen and evaluate the drugs in binding with target proteins 5,[203][204][205][206] .Among all, six drugs including darifenacin (− 7.59 kcal/mol), astemizole (− 7.19), tubocurarine (− 8.26), elacridar (− 7.72), sertindole (− 7.58) and tariquidar (− 8.42) showed higher than − 7 (kcal/mol) docking energy values.The ligand efficiency is a useful metric for lead selection using computational resources 207 .
Ligand efficiency is a way of normalizing the potency and MW of a compound to provide a useful comparison between compounds with a range of MWs and activities.In our docking results, the predicted ligand efficiency (LE) results, showed that all drugs exhibited good LE values, whereas lowest value is considered as most significant compared to other.The comparative results showed that Tubocurarine, Elacridar and Tariquidar exhibited − 0.18 value which was lowest compared to rest of all drugs.Another significant parameter in drug analysis is inhibition constant (Ki).Autodock uses the binding energy (Kcal/mol) to calculate the inhibition constant.The binding energy is the free energy change for the proteininhibitor interaction (ΔG).This is used to determine the inhibition constant (ki) which is, in turn, the dissociation constant (Kd) of the protein-inhibitor complex.In our predicted results, morphine, codeine, quinine, darifenacin,  208 .The intermolecular energy is the attractive intermolecular forces between particles that tend to draw the particles together and evaluates based on minimum energy values.According to AutoDock, the binding energy is the sum of the intermolecular forces acting upon the receptor-ligand complex 209 .
Here, ΔG gauss: attractive term for dispersion of two gaussian functions, ΔGrepulsion: square of the distance if closer than a threshold value, ΔGhbond: ramp function also used for interactions with metal ions, ΔGhydrophobic: ramp function, ΔGtors: proportional to the number of rotatable bonds. (1)

Binding pocket and binding interaction analysis
The active binding site is a cavity on the surface or in the interior of a protein that possesses suitable properties for ligand binding 40 .18 drugs were docked against death domain of MADD and superimposed all docking complexes to check the binding conformations inside the binding pocket.The predicted results showed that all 18 drugs were firmly bind within the active region of MADD however, the binding behavior was deviant with each other's (Fig. 7A, B).Drug-protein binding interaction is significant parameter to better understand the molecular docking results and to deeply understand the conformational behavior 210,211 .Top six drug having docking energy greater than − 7.00 kcal/mol were selected to check the binding conformation behavior and draggability behavior of MADD.The best six drug such as darifenacin (− 7.59), astemizole (− 7.19), tubocurarine (− 8.26), elacridar (− 7.72), sertindole (− 7.58) and tariquidar (− 8.42) exhibited good docking energy values (Kcal/mol) and binding interaction profiles.In darifenacin docking, single hydrogen bond was observed with appropriate binding distance.The amino group (NH 2 ) of darifenacin formed a hydrogen bond with Glu1020 having bond length 2.60 Å within the active region of death domain of MADD structure.Similarly, elacridar and tariquidar also formed hydrogen bonds with Glu1023 having appropriate bond lengths.The nitrogen atom of NH 2 elacridar formed hydrogen bond with Glu1023 having bond length 3.00 Å, whereas tariquidar formed two hydrogen bonds with Glu1023 having bond distances 2.50 and 2.80 Å, respectively.Glutamic acid is an α-amino acid that is used by almost all living beings in the biosynthesis of proteins.Our docking results showed that screened drugs bind with Glu1020 and Glu1023 which may depict their significance in death domain formation and their involvement in the associated signaling pathways.The reported data showed that hydrogen bond between 2.5 and 3.0 Å may considered as standard bond length in ligand docking which strengthen the docked complexes 212 .www.nature.com/scientificreports/Therefore, intermolecular hydrogen bonds have good impact in the formation of drug-protein docking complex and enhance their accuracy.In all our docking analysis, darifenacin, elacridar and tariquidar exhibited comparative hydrogen bond length against the MADD protein.The comparative analysis showed that all drugs, darifenacin, astemizole, tubocurarine, elacridar, sertindole and tariquidar were bind at same position inside the active region of death domain in MADD protein.The common interactions pattern among unique residues also showed the reliability of our docking results.Furthermore, in detail results, darifenacin is surrounded by Glu1020, Tyr1008, Ile1005, Met1004, Gln999, Met997 and Glu1023 residues, whereas elacridar and tariquidar also encompassed similar amino acids such as Leu1027, Leu1030, Met1004, Leu1009, Tyr1008, Glu1020, His1060, and Glu1023, respectively.The other three drugs also showed their binding potential at the same binding regions with different conformational symmetry.The occurrence of common amino acids in all docking complexes also ensure the significance of these amino acid particularly Glu1023 is essential for target binding (Fig. 8).

Molecular dynamic simulation
The best docking complexes were analyzed further through the evaluation of residual flexibility in the target protein by employing MD simulation experiment.The MD simulation study was employed at 100 ns by using Gromacs 5.1.2by generating root mean square deviations & fluctuations (RMSD/F), radius of gyration (Rg) and solvent accessible surface area (SASA) graphs.

Root mean square deviation and fluctuation analysis
The protein backbone behavior in the simulation running time were evaluated through RMSD/F graphs.The RMSD graph results of six docking complexes (darifenacin, astemizole, tubocurarine, elacridar, sertindole and tariquidar) showed the steady and little fluctuated behavior through-out the simulation time.The RMSD graph lines displayed an increasing trend with RMSD values ranging from 0 to 0.1 nm from 0 to 100 ns.
Initially, all the graph lines (red, indigo, brown, green, blue, and yellow) of docked complexes showed an increasing trend with RMSD value 0-0.1 nm from 0 to 10 ns.However, in the same simulation time tariquidar graph line showed much stable behavior as compared to other complexes.From 10 to 20 ns, again fluctuated graphs lines were seen whereas, tariquidar (yellow) remained steady and stable at RMSD value around 0.6 nm.Tubocurarine (brown) displayed highest RMSD value (> 1 nm) at 20 ns as compared to all other graphs lines.Sertindole (blue) line showed upward movement from 0 to 10 ns whereas, from 10 to 20 ns the graph line showed downward movement.However, darifenacin and tubocurarine has been exposed in continuously increasing trend with increased RMSD values (nm).Elacridar (green) and astemizole also showed fluctuated behavior at this simulation time frame.From 20 to 40 ns much stable behavior has been observed in the tariquidar (yellow) and elacridar (green) graphs lines, respectively compared to other graph lines.Darifenacin (red), astemizole (indigo), tubocurarine (brown), elacridar (green), and sertindole (blue) showed high fluctuations in graph lines along with deviated RMSD values.
From 40 to 100 ns, again tariquidar (yellow) remained displayed stable behavior and no fluctuations has been observed in the backbone of MADD protein in the docking complex.Similarly, elacridar (green) also represented similar results with tariquidar and still remined stable behavior in the docking complex.The rest of all other docking complexes showed little fluctuations with steady stable behavior in the simulation time.The RMSD is used to measure the difference between the backbones of a protein from its initial structural conformation to its final position 213 and multiple research data exposed different RMSD values range from 0 to 0.3 nm in their MD simulation analysis 203 .The overall RMSD graphs lines showed that all docked complexes showed fluctuated behavior in the simulation time frame.The generated graphs results showed the stable behavior in the backbone of all protein complexes (Fig. 9).
The RMSF results of all docked complexes dynamically fluctuated from residues N to C terminals.The protein structures are composed of with different structural architecture.There are different small fluctuations through-out the simulation time frame.However, most of protein complexes remain little stable in the RMSF graph.The tariquidar graph line displayed less fluctuations as compared to all other protein complexes which ensure his stable behavior in the docking complex.The comparative results showed that tubocurarine and sertindole were exhibited higher fluctuations peaks, however, the values remain at 0.1 nm in the simulation study.All the results have been displayed in Fig. 10.

Radius of gyration and solvent accessible surface area
The structural compactness of protein was calculated by Rg.The generated results depicted that Rg values of all the docked structures showed little variations from 1.25 to 2.25 nm.Initially, the graph lines were unstable and showed little fluctuations from 0 to 20 ns.The comparative analysis showed that sertindole graph line displayed

Protein-ligand interaction energy
The interaction energy has been calculated from MD trajectories in couple of forms: electrostatic (coulombic) interaction energy and Lennard-Jones interaction energy, with their sum representing the total interaction energy.According to interaction energy analysis tariqudar drug showed the lowest interaction energy − 285.728 followed by elacridar and tobucorarine (− 248.6337 and − 160.0039 respectively).Moreover, the astemizole, sertindole, and darifenacin exhibit high interaction energy as compared to the top (Table 8).Furthermore, the   www.nature.com/scientificreports/graphical depiction of interaction energy of those six drugs has been carried out to see the trajectory changes of all six drugs throughout 100 ns MD simulation (Fig. 13).

Binding poses validation
To analyze the interaction of darifenacin, astemizole, tubocurarine, elacridar, sertindole and tariquidar during 100 ns MD simulation; MD binding Pose validation has been carried out against best selected drugs.Snapshots of all simulated drugs results demonstrated that all drugs remained in the active region of MADD during the 100 ns MD simulation and maintained strong interactions with binding pocket residues.The elacridar manifest the strongest interaction while tubocurarine and sertindole also exhibit single hydrogen bonds and in addition to hydrophobic interactions.Furthermore, astemizole, darifenacin, and tariquidar also exhibits strong hydrophobic interactions.The results demonstrate that the drugs remain bounded to the active pocket of the target protein till 100 ns and block the active site residues (Supplementary Fig. S6).

Conclusion
Novel drug development is time consuming process with relatively high debilitating cost.In present time, drug repositioning is an in-silico approach being employing for drug discovery.In the present research repositioning profiles of known drugs against AD has been explored using shape-based screening, molecular docking pharmacogenomics and MD simulation approaches.Swiss-Similarity results showed 282 drugs were retrieved with donepezil and 351 against galantamine and further evaluated based on similarity scoring values, docking energy values and pharmacogenomics analysis.The detailed pharmacogenomics and extensive data mining showed that three drugs have direct association with AD by targeting different genes.The detailed screening results, showed that darifenacin, astemizole, tubocurarine, elacridar, sertindole and tariquidar drugs exhibited good lead like behavior against AD.Moreover, in MD simulation result tariquidar displayed better stability behavior as compared to rest of other docking complexes drugs with respect to their RMSD, RMSF, SASA and Rg evaluations graphs.Taken together, it has been concluded that tariquidar predicted exhibited better repositioning profiles as compared to other screened FDA approved drugs and may be use in the treatment of AD after in-vitro and clinical assessment in future.
https://doi.org/10.1038/s41598-023-45347-1www.nature.com/scientificreports/Protparam analysis of MADD protein ProtParam (https:// web.expasy.org/ protp aram/) is an online tool which analyse the protein based on various physical and chemical parameters, respectively.The significant parameters include molecular weight (mg/mol), theoretical pI, amino acid composition, atomic composition, extinction coefficient, estimated half-life, instability/aliphatic indexes, and grand average of hydropathicity (GRAVY), respectively.The generated theoretical pI value of MADD protein is calculated by the accumulation of average isotopic masses and pK values of linear amino acids present in the proteins.The prior published research data depicted that proteins are distributed across a wide range of pI values (4.31-11.78) 33.The MADD protein exhibited 5.49 pI value which is comparable with the standard values (4.31-11.78).Moreover, the aliphatic index value also showed the stability and relative volume occupied by aliphatic side chain residues.The predicted physiochemical properties showed the reliability, efficacy, and stability of the MADD protein.The predicted values of both instability and aliphatic indexes 58.81 and 74.96 are justifiable with standard values.The GRAVY value is the sum of hydropathy values of all residues in protein

Figure 2 .
Figure 2. The overall protein structure of MADD protein.The figure showed the three main parts, the overall protein structure, its four core domains and couple of disordered regions.

Figure 5 .
Figure 5. (A-E) SPS analysis of MADD protein to predict their folding and native conformation pattern.

Figure 6 .
Figure 6.(A) Predicted binding pockets represented by graphs lines and scoring values of death domain of MADD protein.(B) the predicted model of death domain and core part has been depicted in purple color in surface format.

Figure 8 .
Figure 8. Binding interaction of best darifenacin, astemizole, tubocurarine, elacridar, sertindole and tariquidar death domain of MADD protein along with all selected drug binding.

Table 1 .
The physiochemical properties of MADD protein.

Table 2 .
The energy evaluation of MADD protein.

energy parameters Predicted scores (Kcal/mol)
acids at positions 58, 60, 174, 696 and 1100, respectively.The graph line remains stable at center value 0 with little deviations range from 2 and − 2. In ES3DC analysis in correlation with Z-score, the generated graph line (yellow)

Table 3 .
Screening of FDA approved drugs using donepezil.Significant values are in [bold].

Table 4 .
Screening of FDA approved drugs using galamtamine against MADD.Significant values are in[bold].

Table 7 .
The predicted docking energy values of selected drugs.

Table 8 .
MD protein-ligand Interaction energy table of all six compounds.