A computational approach to fighting type 1 diabetes by targeting 2C Coxsackie B virus protein with flavonoids

Abstract

Autoimmune diabetes, well-known as type 1 insulin-dependent diabetic mellitus (T1D). T1D is a prolonged condition marked by an inadequate supply of insulin. The lack is brought on by pancreatic cell death and results in hyperglycemia. The immune system, genetic predisposition, and environmental variables are just a few of the many elements that contribute significantly to the pathogenicity of T1D disease. In this study, we test flavonoids against Coxsackie virus protein to cope the type 1 diabetes. After protein target identification we perform molecular docking of flavonoids and selected target (1z8r). then performed the ADMET analysis and select the top compound the base of the docking score and the ADMET test analysis. Following that molecular dynamics simulation was performed up to 300 ns. Root means square deviation, root mean square fluctuation, secondary structure elements, and protein-ligand contacts were calculated as post-analysis of simulation. We further check the binding of the ligand with protein by performing MM-GBSA every 10 ns. Lead compound CID_5280445 was chosen as a possible medication based on analysis. The substance is non-toxic, meets the ADMET and BBB likeness requirements, and has the best interaction energy. This work will assist researchers in developing medicine and testing it as a treatment for Diabetes Mellitus Type 1 brought on by Coxsackie B4 viruses by giving them an understanding of chemicals against these viruses.

Introduction

People with diabetes type 1 have abnormally high blood glucose levels because their systems are unable to manufacture insulin, a hormone that causes the condition. This occurs because the body strikes the tissues that produce insulin in the pancreas, inhibiting it from making insulin in the pancreas, preventing it from making the type of insulin at all. To keep alive, we all need insulin. It serves an important purpose. It allows blood sugar to enter our cells, supplying energy to our bodies. Even if a person develops type 1 diabetes, the body keeps converting carbs from food and beverages into sugar. While glucose flow into the blood, however, there is insufficient insulin to allow it to get stored in the body [1,2]. T1D can manifest itself at any maturity, causing it one of the highly prevalent chronic illnesses in children. Adult-onset type 1 diabetes is more frequent than T1D in childhood and might be misdiagnosed as type 2 diabetes. T1D, which affects 5% to 10% of individuals with diabetes, has progressively grown in incidence and prevalence. A methodical study and extensive analysis found that the frequency of T1D was Fifteen per 100 thousand people, with a global prevalence of 9.5%. Furthermore, there is a large geographical variation in prevalence across the world. Finland and similar Northern European nations have the most reported cases, with rates around 400 times greater than China and Venezuela, which have the lowest known incidences [3–5].

Coxsackievirus B (CVB) is a member of the enterovirus kinds most likely to migrate from the mucous membrane of the intestines to the pancreas in the etiology of diabetes type 1 Mellitus. (T1DM). In the United States, Europe, Asia, Africa, and Australia, a couple of research studies comprising 4,448 and 5,921 individuals, each, verified the importance of the connection between the detection of infection with enteroviral evidence in various human tissue specimens and the likelihood of acquiring islet autoimmunity or T1DM [6,7]. In Type 1 diabetes, a postponed, progressive autoimmune development that may extend for years prior to the onset of apparent illness culminates in selective beta-cell dysfunction [8,9].

CVBs can infect human beta-cells since they are cytolytic viruses. The enterovirus family, which includes the coxsackievirus, is found in the human digestive system. Hepatitis A virus, polioviruses, and hand, foot, and mouth disease (HFMD) are all members of this enterovirus family. This virus transmits quickly from one person to another, generally by direct contact or contact with feces-contaminated surfaces. Because the virus can survive without a host for several days, propagating is very simple [10–12].

In addition, 66 separate serological investigations have identified human enterovirus serotypes. Enteroviruses come in four main varieties. These groups comprise Coxsackie A and B viruses, polioviruses, and echoviruses. The biological characteristics of these classes coincide while being distinct classes. There may be a link between Coxsackie viruses and type 1 diabetes. It has been believed that this theory exists for over 40 years. The nature of this relationship did not, however, support it. Animal models were employed to verify this connection [13]. This viral RNA has been found in the bloodstream of more than fifty percent of T1D affected at the moment of illness start, according to epidemiological data, which revealed a rise in the prevalence of the condition after enterovirus epidemics [14]. These models highlighted the association between T1D and enteroviruses.

A few of the Coxsackie virus B4 isolates from affected with type 1 diabetes with acute onset have been shown to produce diabetes in mice. There are several potential methods by which enteroviruses may initiate or quicken the degenerative processes most important to medical type 1 diabetes. Several enterovirus strains can reproduce in cultured human islet cells, inhibit insulin production, and, in rare instances, result in cell death [15,16].

Coxsackievirus B (CVB) serotypes may have a function in the etiology of type 1 diabetes, according to epidemiological research, although their precise involvement is yet unknown. We have anticipated a CVB1 drug from flavonoids in the current work. Natural remedies are gradually becoming more and more popular because they don’t have any negative side effects in the treatment of brain disorders all over the world. These compounds play a large variety of roles in biological processes. Flavonoids, a class of low molecular weight phenolic compounds, are becoming more and more well-liked due to their wide range of health advantages, their ability to exert a variety of biological properties, including protection from neurological diseases, and their use in nutraceutical, pharmaceutical, medicinal, and cosmetic applications [17,18].

Materials and methods

Preparation of target protein structure and data for docking

Coxsackie B4virus has been detected in recent research employing PCR tests on a variety of diabetes patients. The family of picornaviruses includes coxsackie B4 viruses. These viruses are classified as small RNA viruses because they have a single positive component of RNA. Medicine should be used to address the Coxsackie B4 viral protein, which causes T1D. Coxsackie B4 virus, which kills pancreatic beta cells and causes T1D, is the target of the current in silico project, which aims to find active lead compounds against it using computational methods. The 3D configuration of the targeted protein was retrieved from RCSB PDB by using its specific PDB ID 1Z8R [19]. PDB, the online internet information portal provides access to 3D structural data of macromolecules (proteins, DNA, and RNA) [20]. MODELLER was used for loop refinement [21]. Swiss PDB Viewer [22] and RAMPAGE were used to optimize and minimize the protein crystal structure. RAMPAGE created a Ramachandran Plot that revealed no protein conflicts. The plot also shows which residues are in the favored, allowed, and outlier zones [23]. CASTp (Computed Atlas of Surface Topography of Proteins) predicted protein target binding sites. CASTp 3.0 identifies and quantifies protein topography reliably and thoroughly [24]. Docking was set up with a class of naturally occurring chemicals “flavonoids”. Flavonoids are secondary polyphenolic chemicals found in plants and are a common component of human diets. Flavonoids consist of two phenyl rings and one heterocyclic ring, totaling 15 carbon atoms. Flavonoids’ 2D structures were constructed and minimized. 37 flavonoids compounds were chosen from literature.

Molecular docking

The top 37 compounds from flavonoids were chosen after sorting and screening. Using AutoDock Vina [25], energy dissipated while binding was measured, and protein-ligand interactions were evaluated, after docking these molecules with the receptor. PyMOL [26] was used to create complex receptor and ligand files, whereas BIOVIA Discovery Studio [27] was applied to find interactions in two dimensions.

ADMET analysis

To establish the drug-likeness and toxicity characteristics of compounds, the pkCSM [28] and QikProp developed by Professor William L. Jorgensen [29] were utilized that are reported as essential and valuable tools for evaluation of important druglike descriptors like adsorption, dissemination, and breakdown, elimination, and toxicity (ADMET). These tools are also employed for predicting lead likeness concerning mutagenicity the carcinogenicity. Complete ADMET analysis results are uploaded in S1 Table.

Lead identification

Researchers employed metrics like docking score, ligand-protein interactions, partial coefficient logP, rotatable bonds, rings, Polar Surface Area (PSA), Blood-Brain Barrier, and Ames Toxicity to narrow down the pool of potential inhibitors to a manageable set. 2C Coxsackie B virus protein inhibitor was selected based on their low binding affinities, strong lead-likeness scores, and positive interactions.

MD simulation, PCA and DCCM

Schrödinger LLC’s Desmond program was used for studying 300 ns MD simulations. [30]. The first crucial stage in molecular dynamics modeling, receptor-ligand docking, gives a fixed image of a ligand’s binding location at a protein’s binding site [31]. By including Newton’s classical calculation of action, MD simulations often forecast the status of lead compounds in a biological environment. [32,33].

The Protein Preparation Wizard in Maestro was used to perform preprocessing (optimization and minimization) on the receptor-ligand complex. In this process, steric conflicts, poor contacts, and deformed geometries were eliminated. All structures were built using the System Builder tool, and the OPLS_2005 force field was utilized with the solvent model TIP3P (Intermolecular Interaction Potential 3 Points Transferable), an orthorhombic box [34]. Throughout the simulation period, 310K temperature and 1atm pressure were utilized to imitate physiological circumstances while opposite ions were added to counteract the models, and 0.15M sodium chloride was injected. The models were made looser before the simulation. For examination, frames were saved after every 50 ps, and the binding of protein-ligand was determined over time using RMSD. Using the R package "Bio3D," the principal component analysis (PCA) and dynamic cross-correlation matrix (DCCM) were examined [35]. Simulation trajectory file could be found at this given link: https://drive.google.com/file/d/1hpx9TGYe2nA4z9VOhvwgrvsoBV7Y8UEM/view?usp=sharing

Results and discussion

The 3D structure of the target protein (1Z8R) was obtained from Protein Data Bank. The total structure weight is 18.49 kDa. Fig 1 depicts the protein structure after loop refinement, optimization, and minimization. Ramachandran plot is shown in S1 Fig. The structure’s overall quality was 98 percent, with highly preferred observations. In the plot, all other residues are displayed as circles, while glycine is plotted as triangles and proline as squares. The orange areas are the "favored" areas, the yellow areas are the "allowed" areas, and the white areas are the "disallowed" areas. AutoDock Vina performed the docking of the top hits. ADMET (absorption, distribution, metabolism, excretion, and toxicity) study was accomplished via QikProp and pkCSM. The top 10 compounds are included in Table 1 based on ADMET and docking findings.

[Figure omitted. See PDF.]

[Figure omitted. See PDF.]

In Table 1, mol_MW represents the molecular weight, which should be between 130.0 and 725.0, and donorHB is the projected amount of bonds that the solute would give to water molecules. accepted, the projected amount of bonds the solute would accept from water molecules in an aqueous solution can be a non-integer value with a recommended range of 0.0–6.0. This is because the value is an average across several different configurations. Given that values are calculated as an average over multiple states, they may not all be integers. It operates between 2.0 and 20.0. The Octanol/water partition coefficient, estimated to be in the range of -2.0 to 6.5, is denoted as QPlogPo/w. QPlogHER, Value of the inhibitory concentration (IC50) for the blockade of HERG K+ channels. Negative values below -5 are the cause for alarm. QPPCaco Caco-2 cell permeability prediction, expressed as several nanometers per second. The gut-blood barrier can be mimicked using Caco2 cells. The results of QikProp are for passive transport only. In the range of 0–25, consider it poor, and anything beyond 500 is excellent. QPlogBB Expected brain/blood separation ratio. Dopamine and serotonin, for instance, are CNS-negative because they are too polar to cross the blood-brain barrier, with predicted ranges of—3.0 to—1.2 when using QikProp to predict orally administered drugs. QPlogKhsa and Human serum albumin binding predictions range from -1.5 to 1.5.

Following lead identification, one compound CID: 5280445 was discovered as the most active of all compounds. Fig 2 depicts the best one’s 2D interactions. The properties of the best one are shown in Table 1. Tetrahydroxyflavone luteolin has four hydroxyl groups at positions 3’, 4’, 5, and 7 on its molecular structure. It is believed to function as a key antioxidant, free radical hunter, anti-inflammatory, immunity modulator, and active against several malignancies in the human body [36]. The optimal chemical complex with the protein target was simulated using molecular dynamics simulation for 300 ns. Desmond’s simulated trajectories were analyzed. Root-mean-square-deviation (RMSD) and root-mean-square-fluctuation (RMSF) values, as well as protein-ligand interactions, were determined using MD trajectory analysis.

[Figure omitted. See PDF.]

Fig 3 shows the time-dependent variation in RMSD estimates for C-alpha particles in ligand-bound proteins. The RMSD plot shows that the complexed protein (PDB ID: 1z8r) stabilized at 20 ns. Once the simulation begins, the RMSD stays in the range of 0.5 Angstrom for the rest of the run, which is fine. The experiment validated the general belief that the building was sturdy. Throughout the simulation, there was no significant change in the Ligand Fit to Protein. The RMSD numbers would fluctuate suddenly, sometimes going up and sometimes down. After equilibrium was reached, there was no change in the ligand’s RMSD.

[Figure omitted. See PDF.]

The pink color shows the RMSD of the lead compound and green shows the RMSD of the protein target over time.

Protein dynamics are characterized by PCA (Principal Component Analysis) [37]. Observing collective trajectory motions during MD simulations is a valuable tool. Graph of eigenvalues (protein) vs eigenvector index (eigenmode) for the initial 20 forms of action (NPA022882_1z8r) (Fig 4A). The eigenvalues depict hyperspace eigenvector fluctuations. In simulations, eigenvectors with higher eigenvalues regulate the proteins’ total mobility. The top five eigenvectors in our systems showed dominant movements and had larger eigenvalues (35.4–72.9%) than the other eigenvectors, which had low eigenvalues. More than 50% of all changes were covered by the first three PCs (PC1, PC2, and PC3) that were plotted. According to the Fig 4A plots, PC1 clusters had the largest variability (35.43%), PC2 showed variability (13.69%), and PC3 had the lowest variability (8.99%). As a result of its low variability, PC3 has a more compact structure than PC1 and PC2 and is thought to have a highly steadied protein-ligand binding. Straightforward grouping in the PC subspace showed conformational variations across all groups, with blue exhibiting the most significant mobility, white indicating intermediate movement, and red indicating less flexibility.

[Figure omitted. See PDF.]

(A) Principal Component Analysis eigenvalue plotted versus the percentage of variance (NPA022882_1z8r). The varying areas are displayed in three separate sections. Variations in PC1, PC2, and PC3 add up to 35.43 percent, 13.59 percent, and 8.99 percent, respectively. (B) Complex 5280445_1z8r dynamic cross-correlation map. The residues’ positive and negative correlations are depicted by cyan and purple, respectively.

CID: 5280445 and the 1z8r protein were shown to be significantly correlated with one another, as seen by their high pairwise cross-correlation coefficient value on the cross-correlation map (Fig 4B). Magenta represents anti-correlated residues (-0.4), whereas cyan represents correlated residues (>0.8). It is clear from a large number of pairwise correlated residues between the 1z8r protein and ligand that their binding connection is stable.

Fig 5 illustrates the Residue-wise Root Mean Square Fluctuation (RMSF) of the ligand-coupled protein. Based on MD trajectories, we know that residues with higher peaks are in loop regions or N- and C-terminal regions (Fig 6). The constancy of ligand attachment to the protein is demonstrated by low RMSF estimates of attaching position residues. The secondary structure features of alpha-helices and beta-strands are tracked throughout the simulation (SSE). In the graph below, SSE is plotted against the residue index to display its distribution across the protein structure. Totaling 44.01 percent, it was found that helix made up 2.81 percent, and strand made up 41.20 percent of secondary structure elements.

[Figure omitted. See PDF.]

[Figure omitted. See PDF.]

The alpha helices are represented by the red columns and the beta strands by the blue ones.

In Fig 7, it is clear that hydrogen bonds and hydrophobic interactions constitute most of the important ligand-protein connections established by MD. Hydrogen bonds especially crucial for the amino acids were HIS_5, SER_5, and SER_87. For hydrophobic TYR_3, HIS_21, and TYR_90 are important. The ligand-protein interaction can be monitored over the course of the simulation. In the chart below the contacts and interactions are visualized in a timeline on this figure.

[Figure omitted. See PDF.]

The MMGBSA method is frequently used to evaluate the binding energy of ligands to protein molecules. The influence of additional non-bonded interaction energies as well as the binding free energy of each 5280445_1z8r complex was evaluated. The binding energy of the ligand CID5280445 to 1z8r is -45.5655 kcal/mol. Gbind is governed by non-bonded interactions such as G_bindCoulomb, G_bindPacking, G_bindH_bond, G_bindLipo, and G_bindvdW (Table 2). The S2 Table contains all the MM-GBSA results. The average binding energy was mainly influenced by the G_bindvdW, G_bindLipo, and G_bindCoulomb energies across all types of interactions. The GbindSolvGB and Gbind Covalent energies, on the other hand, made the smallest contributions to the final average binding energies. Additionally, 5280445_1z8r complexes showed stable hydrogen bonds with amino acid residues by their G_bindH_bond interaction values. As a result, the binding energy derived from the docking data was well justified by the MM-GBSA calculations that came from the MD simulation trajectories.

[Figure omitted. See PDF.]

Conclusion

Drug development has been researched extensively since the ability of transdisciplinary strategies to both speed up the process and reduce overall costs. The primary objective of this research was to discover target proteins for 2 Coxsackie B4 viral protein that causes T1D so that a lead drug could be selected for it. To counteract the effects of natural compounds on the 2 Coxsackie B4 viral protein, we chose substances that have this property. An appropriate natural inhibitor, identified from flavonoids CID: 5280445, blocks the action of 1z8r at its receptor. We reasoned that this material might act as a beginning point for the progress of a medication that targets Diabetes Type 1 (T1D) selectively without affecting other cellular processes. These results will be useful to researchers and may lead to the progress of a new medicine for the treatment of T1D.

Supporting information

S1 Table. Complete results of ADMET analysis.

https://doi.org/10.1371/journal.pone.0290576.s001

(CSV)

S2 Table. MM-GBSA binding energy calculation of bonded and non-bonded interactions of CID_5280445 with 1z8r after every 10 ns from MD Simulation trajectories.

https://doi.org/10.1371/journal.pone.0290576.s002

(CSV)

S1 Fig.

https://doi.org/10.1371/journal.pone.0290576.s003

(PNG)

Citation: Ullah S, Zheng Z, Rahman W, Ullah F, Ullah A, Iqbal MN, et al. (2023) A computational approach to fighting type 1 diabetes by targeting 2C Coxsackie B virus protein with flavonoids. PLoS ONE 18(8): e0290576. https://doi.org/10.1371/journal.pone.0290576

About the Authors:

Shahid Ullah

Contributed equally to this work with: Shahid Ullah, Zilong Zheng

Roles: Supervision, Writing – original draft, Writing – review & editing

E-mail: [email protected] (SU); [email protected] (TG)

Affiliation: S Khan Lab Mardan, Khyber Pakhtunkhwa, Pakistan

ORICD: https://orcid.org/0000-0001-6694-5590

Zilong Zheng

Contributed equally to this work with: Shahid Ullah, Zilong Zheng

Roles: Methodology, Validation, Visualization

Affiliation: Big Data Center, The Seventh Affiliated Hospital of Sun Yat-sen University, Shenzhen, P. R. China

Wajeeha Rahman

Roles: Data curation, Formal analysis, Methodology, Validation, Visualization

Affiliation: S Khan Lab Mardan, Khyber Pakhtunkhwa, Pakistan

Farhan Ullah

Roles: Data curation, Validation, Visualization

Affiliation: S Khan Lab Mardan, Khyber Pakhtunkhwa, Pakistan

Anees Ullah

Roles: Data curation, Formal analysis, Validation, Visualization

Affiliation: S Khan Lab Mardan, Khyber Pakhtunkhwa, Pakistan

Muhammad Nasir Iqbal

Roles: Data curation, Validation, Visualization

Affiliation: Department of Bioinformatics, Institute of Biochemistry, Biotechnology and Bioinformatics, The Islamia University of Bahawalpur, Bahawalpur, Pakistan

Naveed Iqbal

Roles: Data curation, Methodology, Validation

Affiliation: Department of Bioinformatics, Institute of Biochemistry, Biotechnology and Bioinformatics, The Islamia University of Bahawalpur, Bahawalpur, Pakistan

Tianshun Gao

Roles: Supervision, Visualization, Writing – review & editing

E-mail: [email protected] (SU); [email protected] (TG)

Affiliation: Big Data Center, The Seventh Affiliated Hospital of Sun Yat-sen University, Shenzhen, P. R. China

References

1. Roep BO. The role of T-cells in the pathogenesis of Type 1 diabetes: from cause to cure. Diabetologia. 2003;46(3):305–21. pmid:12687328

2. Ehrmann D, Kulzer B, Roos T, Haak T, Al-Khatib M, Hermanns N. Risk factors and prevention strategies for diabetic ketoacidosis in people with established type 1 diabetes. Lancet Diabetes Endocrinol. 2020;8(5):436–46. pmid:32333879

3. Mobasseri M, Shirmohammadi M, Amiri T, Vahed N, Hosseini Fard H, Ghojazadeh M. Prevalence and incidence of type 1 diabetes in the world: a systematic review and meta-analysis. Health Promot Perspect. 2020;10(2):98–115. pmid:32296622

4. Leslie RD, Evans-Molina C, Freund-Brown J, Buzzetti R, Dabelea D, Gillespie KM, et al. Adult-Onset Type 1 Diabetes: Current Understanding and Challenges. Diabetes Care. 2021;44(11):2449–56. pmid:34670785

5. Holt RIG, DeVries JH, Hess-Fischl A, Hirsch IB, Kirkman MS, Klupa T, et al. Correction to: The management of type 1 diabetes in adults. A consensus report by the American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD). Diabetologia. 2022;65(1):255.

6. Yeung WC, Rawlinson WD, Craig ME. Enterovirus infection and type 1 diabetes mellitus: systematic review and meta-analysis of observational molecular studies. BMJ. 2011;342:d35. pmid:21292721

7. Jaidane H, Sauter P, Sane F, Goffard A, Gharbi J, Hober D. Enteroviruses and type 1 diabetes: towards a better understanding of the relationship. Rev Med Virol. 2010;20(5):265–80. pmid:20629044

8. Kobayashi T, Tanaka S, Aida K. Unique pathological changes in the pancreas of fulminant type 1 diabetes. Diabetol Int. 2020;11(4):323–8. pmid:33088638

9. Tanaka S, Aida K, Nishida Y, Kobayashi T. Pathophysiological mechanisms involving aggressive islet cell destruction in fulminant type 1 diabetes. Endocr J. 2013;60(7):837–45. pmid:23774118

10. Nekoua MP, Mercier A, Vergez I, Morvan C, Mbani CJ, Sane F, et al. [Coxsackievirus B infection and pathogenesis of type 1 diabetes]. Virologie (Montrouge). 2022;26(6):415–30.

11. Emekdas G, Rota S, Kustimur S, Kocabeyoglu O. [Antibody levels against coxsackie B viruses in patients with type 1 diabetes mellitus]. Mikrobiyol Bul. 1992;26(2):116–20.

12. Nekoua MP, Alidjinou EK, Hober D. Persistent coxsackievirus B infection and pathogenesis of type 1 diabetes mellitus. Nat Rev Endocrinol. 2022;18(8):503–16. pmid:35650334

13. Banatvala JE. Insulin-dependent (juvenile-onset, type 1) diabetes mellitus Coxsackie B viruses revisited. Prog Med Virol. 1987;34:33–54.

14. Yin H, Berg AK, Tuvemo T, Frisk G. Enterovirus RNA is found in peripheral blood mononuclear cells in a majority of type 1 diabetic children at onset. Diabetes. 2002;51(6):1964–71. pmid:12031987

15. Hyoty H, Hiltunen M, Lonnrot M. Enterovirus infections and insulin dependent diabetes mellitus—evidence for causality. Clin Diagn Virol. 1998;9(2–3):77–84. pmid:9645988

16. Roivainen M, Ylipaasto P, Savolainen C, Galama J, Hovi T, Otonkoski T. Functional impairment and killing of human beta cells by enteroviruses: the capacity is shared by a wide range of serotypes, but the extent is a characteristic of individual virus strains. Diabetologia. 2002;45(5):693–702. pmid:12107750

17. Atanasov AG, Zotchev SB, Dirsch VM, Orhan IE, Banach M, Rollinger JM, et al. Natural products in drug discovery: advances and opportunities. Nature Reviews Drug Discovery. 2021;20(3):200–16. pmid:33510482

18. Panche AN, Diwan AD, Chandra SR. Flavonoids: an overview. Journal of Nutritional Science. 2016;5:e47. pmid:28620474

19. Baxter NJ, Roetzer A, Liebig HD, Sedelnikova SE, Hounslow AM, Skern T, et al. Structure and dynamics of coxsackievirus B4 2A proteinase, an enyzme involved in the etiology of heart disease. J Virol. 2006;80(3):1451–62. pmid:16415022

20. Berman HM, Westbrook J, Feng Z, Gilliland G, Bhat TN, Weissig H, et al. The Protein Data Bank. Nucleic Acids Res. 2000;28(1):235–42. pmid:10592235

21. Eswar N, Webb B, Marti-Renom MA, Madhusudhan MS, Eramian D, Shen MY, et al. Comparative protein structure modeling using Modeller. Curr Protoc Bioinformatics. 2006;Chapter 5:Unit-5 6. pmid:18428767

22. Guex N, Peitsch MC. SWISS-MODEL and the Swiss-PdbViewer: an environment for comparative protein modeling. Electrophoresis. 1997;18(15):2714–23. pmid:9504803

23. Ho BK, Brasseur R. The Ramachandran plots of glycine and pre-proline. BMC Struct Biol. 2005;5:14. pmid:16105172

24. Tian W, Chen C, Lei X, Zhao J, Liang J. CASTp 3.0: computed atlas of surface topography of proteins. Nucleic Acids Res. 2018;46(W1):W363–W7. pmid:29860391

25. Trott O, Olson AJ. AutoDock Vina: improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading. J Comput Chem. 2010;31(2):455–61. pmid:19499576

26. Mura C, McCrimmon CM, Vertrees J, Sawaya MR. An introduction to biomolecular graphics. PLoS Comput Biol. 2010;6(8). pmid:20865174

27. Systèmes D. BIOVIA Discovery Studio. San Diego. 2022.

28. Pires DE, Blundell TL, Ascher DB. pkCSM: Predicting Small-Molecule Pharmacokinetic and Toxicity Properties Using Graph-Based Signatures. J Med Chem. 2015;58(9):4066–72. pmid:25860834

29. Laoui A, Polyakov VR. Web services as applications’ integration tool: QikProp case study. J Comput Chem. 2011;32(9):1944–51. pmid:21455963

30. Bowers KJaC David E. and Xu Huafeng and Dror , Ron O. and Eastwood , Michael P. and Gregersen , Brent A. and Klepeis , et al. Scalable Algorithms for Molecular Dynamics Simulations on Commodity Clusters: IEEE; 2006. 43– p.

31. Ferreira LG, Dos Santos RN, Oliva G, Andricopulo AD. Molecular docking and structure-based drug design strategies. Molecules. 2015;20(7):13384–421. pmid:26205061

32. Hildebrand PW, Rose AS, Tiemann JKS. Bringing Molecular Dynamics Simulation Data into View. Trends Biochem Sci. 2019;44(11):902–13. pmid:31301982

33. Rasheed MA, Iqbal MN, Saddick S, Ali I, Khan FS, Kanwal S, et al. Identification of Lead Compounds against Scm (fms10) in Enterococcus faecium Using Computer Aided Drug Designing. Life (Basel). 2021;11(2). pmid:33494233

34. Shivakumar D, Williams J, Wu Y, Damm W, Shelley J, Sherman W. Prediction of Absolute Solvation Free Energies using Molecular Dynamics Free Energy Perturbation and the OPLS Force Field. Journal of Chemical Theory and Computation. 2010;6(5):1509–19. pmid:26615687

35. Grant BJ, Skjaerven L, Yao XQ. The Bio3D packages for structural bioinformatics. Protein Sci. 2021;30(1):20–30. pmid:32734663

36. Imran M, Rauf A, Abu-Izneid T, Nadeem M, Shariati MA, Khan IA, et al. Luteolin, a flavonoid, as an anticancer agent: A review. Biomed Pharmacother. 2019;112:108612. pmid:30798142

37. David CC, Jacobs DJ. Principal component analysis: a method for determining the essential dynamics of proteins. Methods Mol Biol. 2014;1084:193–226. pmid:24061923