Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Open Access

Peer-reviewed

Research Article

Cheminformatics-based screening and evaluation of phytochemicals as CDK2 inhibitors in colorectal cancer therapy

  • Nowshin Tabassum,

    Roles Data curation, Formal analysis, Methodology, Validation, Visualization, Writing – original draft, Writing – review & editing

    Affiliations Bioinformatics Laboratory (BioLab), Noakhali, Bangladesh, Department of Microbiology, Noakhali Science and Technology University, Noakhali, Bangladesh

    ⨯
  • Md. Ekram Hossan,

    Roles Data curation, Formal analysis, Methodology, Validation, Visualization, Writing – original draft, Writing – review & editing

    Affiliations Bioinformatics Laboratory (BioLab), Noakhali, Bangladesh, Department of Microbiology, Noakhali Science and Technology University, Noakhali, Bangladesh

    ⨯
  • Md. Mujahidul Islam,

    Roles Data curation, Formal analysis, Methodology, Visualization, Writing – original draft

    Affiliations Bioinformatics Laboratory (BioLab), Noakhali, Bangladesh, Department of Biochemistry and Molecular Biology, University of Chittagong, Chittagong, Bangladesh

    ⨯
  • Mahmudul Hasan,

    Roles Data curation, Formal analysis, Methodology, Validation, Visualization, Writing – original draft, Writing – review & editing

    Affiliations Bioinformatics Laboratory (BioLab), Noakhali, Bangladesh, Department of Microbiology, Noakhali Science and Technology University, Noakhali, Bangladesh

    ⨯
  • Md Salamoon Islam,

    Roles Writing – original draft

    Affiliations Bioinformatics Laboratory (BioLab), Noakhali, Bangladesh, Department of Chemistry, Jashore University of Science and Technology, Jashore, Bangladesh

    ⨯
  • Sunjida Masud,

    Roles Writing – original draft

    Affiliations Bioinformatics Laboratory (BioLab), Noakhali, Bangladesh, Department of Microbiology, Noakhali Science and Technology University, Noakhali, Bangladesh

    ⨯
  • Firoz Ahmed ,

    Roles Investigation, Resources, Software, Supervision, Validation, Writing – review & editing

    noimul.microbiology.nstu@gmail.com (NHS); firoz@nstu.edu.bd (FA)

    Affiliations Bioinformatics Laboratory (BioLab), Noakhali, Bangladesh, Department of Microbiology, Noakhali Science and Technology University, Noakhali, Bangladesh

    ⨯
  • Noimul Hasan Siddiquee

    Roles Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Resources, Software, Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing

    noimul.microbiology.nstu@gmail.com (NHS); firoz@nstu.edu.bd (FA)

    Affiliations Bioinformatics Laboratory (BioLab), Noakhali, Bangladesh, Department of Microbiology, Noakhali Science and Technology University, Noakhali, Bangladesh

    ⨯

Abstract

Colorectal cancer (CRC) poses a significant global health issue. It ranks as the third most common type of cancer and the second leading cause of cancer-related deaths. Among the molecular factors driving its progression, cyclin-dependent kinase 2 (CDK2) plays a key role. CDK2 is a protein kinase essential for regulating the cell cycle, and its dysregulation is implicated in the development of various cancers, notably CRC. Fruquintinib is an already available drug against CRC. However, this study is being performed in search of better drug-like compounds. Some studies have shown that phytochemicals are less toxic and have fewer adverse effects than commercially available medications. With the vision of detecting CDK2 inhibitors, phytochemicals with anticancer activity can be used as alternatives to develop the drug candidate. Cheminformatics-based analysis is used for this purpose. This approach includes molecular docking, adsorption, distribution, metabolism, excretion/toxicity (ADME/T), post-docking molecular mechanism generalized born surface area (MM-GBSA), structural activity relationship (SAR), frontier molecular orbital (FMO), and molecular dynamics (MD) simulations. Molecular docking was employed to determine the binding strength of 4433 phytochemicals with anti-cancer properties sourced from the IMPPAT database. The top five candidates, CIDs-135438111, 6474893, 44257567, 10469828, and 353825, were selected based on their docking scores. Later, three lead compounds, CIDs-6474893, 10469828, and 135438111, were finalized depending on their favorable ADME/T profiles. All three selected pharmaceuticals demonstrated excellent post-dock MM-GBSA scores and HOMO-LUMO energy gaps, which served as confirmation of their efficacy and safety. The SAR analysis also revealed anti-mutagenic, antineoplastic, and apoptosis-inducing properties of the compounds. Finally, the rigidity of the protein-ligand complex structures was verified by MD simulations. Overall, the study suggests these three phytochemicals exhibit stronger binding and better pharmacological profiles than the control (fruquintinib), offering a promising direction for CRC treatment development.

Citation: Tabassum N, Hossan ME, Islam MM, Hasan M, Islam MS, Masud S, et al. (2025) Cheminformatics-based screening and evaluation of phytochemicals as CDK2 inhibitors in colorectal cancer therapy. PLoS One 20(9): e0331438. https://doi.org/10.1371/journal.pone.0331438

Editor: Ahmed A. Al-Karmalawy, University of Mashreq, IRAQ

Received: July 19, 2025; Accepted: August 17, 2025; Published: September 3, 2025

Copyright: © 2025 Tabassum et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: All relevant data are within the manuscript and its Supporting Information files.

Funding: The author(s) received no specific funding for this work.

Competing interests: The authors have declared that no competing interests exist.

1. Introduction

Cancer ranks as the second leading cause of death globally [1]. CRC stands as the fourth predominant malignancy and the third most significant contributor of cancer-related deaths worldwide [2]. CRC affects individuals of all ages, though a higher incidence is observed in those over 50 [3,4]. Contrary to popular belief, the lifetime average risk of colorectal cancer (CRC) is only 5% worldwide, and studies reveal that low and middle-income nations would see a 70% increase in CRC incidence by 2030 [2,5]. Abnormal mitosis and proliferation of colon cells are the causes of CRC [6], with genetic factors [7], lifestyle, habits, and diet being major contributors [8]. Risk factors encompass obesity, smoking, drinking alcohol, and diets rich in fat [9]. In contrast, genetic factors, like familial adenomatous polyposis coli (FAP) and hereditary non-polyposis colon cancer (HNPCC, Lynch syndrome), also exert considerable influence [10]. Although colorectal cancer may be asymptomatic in its initial phases, symptoms including rectal bleeding, decreased frequency of bowel movements, rectal bleeding, abdominal pain, weakness, and weight loss manifest as the disease progresses [11].

Colorectal cancer treatment has conventionally depended on chemotherapy, utilizing agents such as fluoropyrimidine and oxaliplatin, along with the surgical excision of tumors and metastases. In some instances, neoadjuvant or adjuvant chemotherapy or radiation may be administered before or after surgery to shrink and stabilize the tumor optimally [12]. However, existing therapeutic strategies are linked to significant side effects and are frequently worsened by drug resistance [13,14]. Consequently, the pursuit of novel, more precise treatments persists as researchers endeavour to create medicines that are more targeted and efficacious [15]. Recent years have witnessed significant advancements in molecular targeting, particularly targeting the vascular endothelial growth factor (VEGF), CDKs, and epidermal growth factor receptor (EGFR) [10].

The formation of new blood vessels associated with tumor growth is impeded by a VEGFR inhibitor known as fruquintinib [16]. Fruquintinib has been utilized to treat metastatic colorectal cancer (mCRC) by targeting the tyrosine kinase linked to VEGFR-1, VEGFR-2, and VEGFR-3, respectively [17]. Due to its strong suppression of VEGFR-1, 2, and 3 receptors, fruquintinib inhibits angiogenesis, which hinders tumor growth, and it inhibits lymphangiogenesis, which hinders tumor dispersion [18]. In vivo, oral fruquintinib treatment significantly reduced VEGF-induced VEGFR2 phosphorylation [19]. Through anti-angiogenesis, fruquintinib inhibited tumor growth in a patient-derived tumor xenograft model while preserving a favorable dose-response relationship [20]. Fruquintinib and anti-PD-1 (Programmed death receptor-1) therapy together may improve the antitumor immune response in microsatellite-stable colorectal cancer (MSS CRC) [21]. A monoclonal antibody that targets PD-1 is called sintilimab. When used in mice models with MC38 or CT26 xenograft tumors, which stand in for MSS and microsatellite-instability colorectal cancer, fruquintinib and sintilimab significantly inhibited tumor growth and extended survival time when compared to either drug alone [22].

In recent years, natural products like phytochemicals have gained attention for cancer therapy due to their distinctive structures, reduced toxicity, cost efficiency, especially in impoverished countries, and capacity to target several cancer-associated signaling pathways [23,24]. More than 60% of anticancer drugs originate from natural sources, and studies have shown that phytochemicals can disrupt cancer growth by altering pathways such as cellular proliferation, apoptosis, and angiogenesis [25,26]. Withania somnifera’s phytochemicals show promising anticancer properties with fewer side effects than synthetic drugs. Targeting cancer stem cells improves treatment outcomes and reduces recurrence [27]. The oxidative stress and inflammation that are associated with cancer growth are alleviated by their antioxidant and anti-inflammatory properties, which also influence cancer mechanisms such as apoptosis [27–29]. Natural chemicals such as terpenoids, alkaloids, and flavonoids present varied structures and modes of action for the identification of anticancer drugs. They can trigger apoptosis, impede cell proliferation, halt cell cycles, limit metastasis, and possess chemopreventive capabilities [30]. Additionally, they exhibit resistance mitigation, rendering them potent anticancer agents, and display various biological activities [31]. Phytochemicals have low toxicity, multi-target mechanisms, synergistic interactions [32], varied chemical structures, provide specific mechanisms [33], biological activity, diverse methods of action, bioavailability, accessibility, and synergistic effects, rendering them safer, more sustainable, and perhaps mitigating medication resistance [34]. The significance of phytochemicals in chemosensitization offers a possible approach for overcoming drug resistance, an essential barrier in colorectal cancer treatment [35].

CDK-2 (Fig 1) is essential in cell cycle regulation and is overexpressed in colorectal cancer, among other malignancies, rendering it an attractive target for drug development [36,37]. The phosphorylation of targets implicated in DNA replication is regulated by CDK2 in complex with cyclin A during the S phase of the cell cycle, as illustrated in Fig 2 [38]. The Rb (retinoblastoma) protein is a protein that inhibits tumor development and is responsible for the regulation of the cell cycle, differentiation, and apoptosis [39]. Phosphorylation of the retinoblastoma (Rb) protein is a key factor in regulating the cell cycle [40]. The Rb protein has many phosphorylation sites that are targeted by kinases such as cyclinD/cdk4, cyclinE/cdk2, and cyclinA/cdk2 [39]. The overexpression of cyclins or the loss of expression of CDKIs is a significant factor in the high phosphorylation of the tumor suppressor Retinoblastoma (Rb) protein in cancer cells [41].

thumbnail
Download:
Fig 1. CDK2 protein (PDB ID-6GUE).

https://doi.org/10.1371/journal.pone.0331438.g001

thumbnail
Download:
Fig 2. The CDK/cyclin complexes and the cell cycle phases.

https://doi.org/10.1371/journal.pone.0331438.g002

Although CDK-1 can substitute for CDK-2 in normal cells, CDK-2 becomes essential in malignant cells, facilitating cancer growth [36,42]. In the ongoing research, computational methodologies, including molecular docking, MD simulations, MM-GBSA, and ADME/T, have accelerated the discovery of CDK-2 inhibitors, providing a pathway to more selective and effective therapies [43]. Inhibitors such as sorafenib and aspirin have demonstrated potential in downregulating CDK-2 [44,45]; nevertheless, concerns regarding specificity and toxicity persist, requiring further research into more targeted therapies [46].

This study provides a thorough examination of CDK2 inhibitors, with a specific emphasis on the elucidation of their molecular mechanisms of action, pharmacological efficacy, and toxicity profiles. By leveraging cheminformatics approaches, such as molecular docking and MD simulations, this research identifies and characterizes promising lead compounds with high binding affinity and favorable pharmacokinetic properties. These findings contribute to the structure-based rational drug design pipeline, facilitating the development of selective and potent CDK2 inhibitors as potential therapeutic agents, particularly for the treatment of CDK2-driven malignancies such as colorectal cancer.

2. Methods and materials

The whole experimental framework is graphically presented in Fig 3.

thumbnail
Download:
Fig 3. Workflow of the identification of phytochemicals that can effectively inhibit overexpressed CDK2 through the cheminformatics method.

https://doi.org/10.1371/journal.pone.0331438.g003

2.1 Sourcing and pre-docking preparation of protein and ligand

The crystal X-ray diffraction structure of CDK2 (PDB ID-6GUE), a protein that regulates the cell cycle in humans, was obtained from the RCSB Protein Data Bank (https://www.rcsb.org/) and shown in three dimensions. The X-ray diffraction method was employed to assess the structure of CDK2 with a resolution of 1.99 Å (generally referred to as < 2.5 Å). The R-value free was 0.234 [42]. The default Protein Preparation Wizard option of Schrödinger 2023–1 was used to remove chains (B, C, and D), ligands, hetero atoms, and water molecules as they often interrupt the docking process (Fig 4). The chain A-associated ligand was kept for the crystalline ligand-based grid box generation. The OPLS4 (Optimized Potentials for Liquid Simulations, version 4 extended) force field was utilized for prediction accuracy. The default parameter of the protein preparation wizard is used for assigning proper bond ordering, adding the H atom, and adding missing amino acid residues. Now, chain A is ready to be used effectively for screening and finding a lead compound for treating colorectal cancer.

thumbnail
Download:
Fig 4. Chain A from the native CDK2 protein was isolated for molecular docking.

https://doi.org/10.1371/journal.pone.0331438.g004

The 3D conformation of the reference molecule, fruquintinib (CID-44480399), was obtained in SDF format from the PubChem database (https://pubchem.ncbi.nlm.nih.gov/). Fruquintinib is a therapeutic agent that inhibits VEGF for treating CRC [47]. A total of 4433 ligands were extracted from the IMPPAT database (https://cb.imsc.res.in/imppat/). Previous studies have shown that most plants have therapeutic action against human cervical, skin, colon, and lung cancer (Supplementary S1). Further utilization processes may include ligands prepared by the LigPrep module of Maestro 13.5, applying the OPLS4 force field, and decreasing ionization state by Epik at PH 7.

2.2 Molecular docking analysis

A computational approach to establish stable complexes between larger biomolecules (such as proteins, DNA, or RNA) and smaller molecules (such as ligands) is referred to as molecular docking [48]. The Schrödinger Protein Preparation Wizard was used to construct CDK2 (6GUE, chain A) to anticipate the best ligand-protein match via molecular docking studies. A grid box is created for the receptor, with the native ligand in the middle. The virtual screening was conducted by docking 4,433 compounds and controls with the target protein by using GLIDE available in Maestro v13.5 of Schrödinger, which led to the identification of the ligand exhibiting the highest docking score. The OPLS4 force field was employed in the docking process using standard precision mode. The Maestro viewer revealed the interacting chemical bonds that connect ligands and amino acid residues in the protein.

2.3 Pharmacokinetics and toxicity prediction

The interaction of pharmacokinetics, toxicity, and potency is essential for medication efficacy. The ADME/T profile of a compound ensures its safe and effective delivery to the target. Preliminary ADME/T assessment has diminished clinical trial failures due to inadequate pharmacological characteristics [49]. The control, along with the five top-scoring compounds from the docking analysis, was subjected to an in-depth evaluation of ADME properties and toxicity predictions to assess their pharmacokinetic and safety profiles.

To evaluate the ADME, physicochemical, drug-likeness, and pharmacokinetic properties of the chosen ligand compounds, the SwissADME online tool (http://www.swissadme.ch) was used [49]. Since small compounds have the potential to harm many human organs via cytotoxicity, carcinogenicity, hepatotoxicity, immunotoxicity, and mutagenicity, it is essential to assess their potential toxicity. Predicting toxicity is thus crucial to the medication development process. ProTox 3.0 (https://tox-new.charite.de/) was used in this investigation to estimate the toxicity profiles of the selected substances [50].

2.4 Post-docking MM-GBSA calculation

The MM-GBSA method was used to analyze the free binding energies of the protein and ligand complexes [51]. ΔGbind = GRL − GR − GL is the formula used to represent the binding energy of the ligand (L) to the receptor (R), which results in the complex (RL) [52]; a larger negative score denotes a bigger and longer binding energy [53]. The lead compounds with a lower docking score and those that successfully satisfied the ADME/T criteria were selected for further analysis using MM-GBSA computations. Maestro’s Prime MM-GBSA v3.00 software was used to calculate the MM-GBSA score. An indication for assessing the effectiveness of recently screened medications is the MM-GBSA score in conjunction with docking scores.

2.5 SAR analysis

The PASS online web tool (http://way2drug.com/PassOnline/predict.phpp) was used to evaluate the anticancer potential activity spectrum of the selected phytochemical substances and the control. The PASS program uses a database of more than 4,000 attributes to find bioactive chemicals with high prediction reliability, including drug-related and non-drug-related activities. The appropriate prediction findings are then expressed using both the probability of activity (Pa) and the probability of inactivity (Pi), which range from 0.00 to 1.00. A molecule must meet the requirement that Pa > Pi, which means that the total of Pa and Pi is equal to 1, to be deemed not completely inactive or active simultaneously [54]. The findings of the PASS prediction were used and understood in a flexible way, demonstrating that (i) when Pa > 0.7, there is a high probability of detecting the activity empirically, (ii) if 0.5 < Pa < 0.7, there is a low probability of detecting the activity empirically, even though the chemical probably isn’t as comparable to existing pharmaceuticals, and (iii) if Pa < 0.5, there is a lower probability of detecting the activity empirically, but there is a probability of detecting a structurally similar pharmaceutical agent [55].

2.6 FMO analysis

The FMO analysis is regarded as a comprehensive method for examining the optoelectronic characteristics of a molecule [56]. The electronic properties of the selected compounds CIDs-6474893, 10469828, and 135438111 were examined by FMO analysis [57], focusing on the Highest Occupied Molecular Orbital (HOMO) and the Lowest Unoccupied Molecular Orbital (LUMO) [58]. The LUMO serves as an electron acceptor in molecular interactions, whereas the HOMO functions as an electron donor [59]. The Gaussian09 software [60] package was employed to conduct all calculations, and the output of these calculations was visualized by using the GaussView program [61]. The computations used the 6-31G basis set [62] and the Density Functional Theory (DFT) [63] approach with the restricted B3LYP (RB3LYP) [64] hybrid function. The HOMO–LUMO energy gap (ΔE), molecular hardness (η), and molecular softness (S) were determined to find out how reactive and stable the molecules are [65]. Parr and Pearson’s interpretation of DFT and Koopmans’ theorem was used to come up with these numbers, which link electron affinity (A) and ionization potential (I) to the energies of LUMO and HOMO orbitals [66,67]. The following equations were applied:

HOMO-LUMO energy gap, ΔE = (ELUMO – EHOMO),

The energy of the HOMO is denoted as EHOMO, and that of the LUMO as ELUMO.

Hardness, η = (I-A)/2

Softness, S = 1/η

A smaller ΔE value indicates greater reactivity and lower kinetic stability, while a larger ΔE value indicates greater stability and reduced reactivity [68,69]. Hardness (η) and softness (S) are global reactivity descriptors computed with the energies of the frontier molecular orbitals [65]. A more stable and less reactive molecule results from a larger Hardness (η) value [70]. Conversely, softness (S) indicates the capacity of the molecule to allow electrical alterations [71,72]. A reduced hardness (i.e., increased softness) indicates enhanced chemical reactivity and polarizability [73].

2.7 MD simulation

MD simulation is a highly effective approach for clarifying the interactions between potential pharmaceuticals and their biological targets [74]. Ligands and proteins may not regularly attain precise binding poses within the receptor pocket. Molecular Dynamics (MD) simulations can refine binding poses by employing force fields to assess dynamic interactions, facilitating modifications in ligand orientation to improve affinity and specificity. This procedure is essential for optimizing ligand design and enhancing therapeutic efficacy [75]. The Desmond module from Schrödinger was employed in a Linux environment to simulate molecular dynamics for 200 ns for this investigation [76]. The simple point-charge (SPC) water model was used to accurately characterize the membrane environment to guarantee appropriate electrical permittivity and water density. Na+ and Cl+ ions were used to neutralize the system and achieve a physiological salt concentration of 0.15 M. To keep the volume of a 10 × 10 × 10 ų orthorhombic box constant, periodic boundary conditions were created. At 300 K and 1.01325 bar with 1.2 energy, molecular dynamics (MD) simulations were carried out in NPT ensemble conditions. Trajectories were recorded at 200 picosecond intervals to define interactions in the solvated protein-ligand system using the OPLS4 Force Field. By setting these settings, it was possible to accurately simulate the biological system and gain necessary knowledge about how proteins and ligands interact.

3. Results

3.1 Molecular docking analysis

Molecular docking was used using 4433 drug-like compounds and the control drug fruquintinib (CID-44480399). Among the top five compounds with the highest docking energy, three had higher docking energy than fruquintinib (control). The top five compounds, which have a greater binding affinity towards the target, are CID-135438111 (IMPHY008184), CID-6474893 (IMPHY007097), CID-44257567 (IMPHY013402), CID-10469828 (IMPHY013402), and CID-353825 (IMPHY011367), compared to the control CID-44480399, as shown in Fig 5. Their docking scores are −8.037, −9.271, −9.255, −9.255, −9.218, and −9.176 kcal/mol, respectively (Table 1).

thumbnail
Download:
Table 1. Identification, docking energy score, source, and therapeutic applications of the top five chosen compounds and control (fruquintinib).

https://doi.org/10.1371/journal.pone.0331438.t001

thumbnail
Download:
Fig 5. Molecular docking energy scores (A), 3D structures of the five best docked compounds (CID-135438111, 6474893, 10469828, 44257567, and 353825) and the control (CID-44480399) (B).

https://doi.org/10.1371/journal.pone.0331438.g005

3.2 Pharmacokinetics and toxicity prediction

The drug’s pharmacokinetic profile can be improved by employing ADME properties, which are essential for comprehending the drug’s overall pharmacokinetics. Based on the docking scores of 4433 compounds, the top five phytochemical compounds (CID-135438111, 6474893, 44257567, 10469828, and 353825) and the control drug (CID-44480399) undergo ADME, pharmacokinetics, and toxicity properties analysis. From the inquiry, it was observed that three phytochemicals (CID-6474893, CID-10469828, and CID-135438111) showed druggable ADME, toxicity characteristics, and no Lipinski violations. Conversely, the control ligand (CID-44480399) exhibited active toxicity in various toxic properties, in comparison to the three lead compounds. Thus, the above three phytochemical compounds were selected for further analysis. The pharmacokinetics and toxicity properties of the three lead compounds (CIDs-6474893, 10469828, and 135438111) and the control (CID-44480399) are shown in Figs 6 and 7, respectively, and additional data are provided in Supplementary S2.

thumbnail
Download:
Fig 6. ADME properties of three lead compounds (CIDs-6474893, 10469828, and 135438111) and the control (CID-44480399).

https://doi.org/10.1371/journal.pone.0331438.g006

thumbnail
Download:
Fig 7. Toxicity properties of the three lead compounds (CIDs-6474893, 10469828, and 135438111) and the control (CID-44480399).

https://doi.org/10.1371/journal.pone.0331438.g007

3.3 Protein-ligand interaction analysis

Based on binding affinity and ADME/T properties, three lead compounds, CIDs-6474893, 10469828, 135438111, and control CID-44480399, were chosen for further downstream analysis. Here, the maestro module of the Schrödinger suite was employed for visualizing molecular interactions. The interaction of amino acids with the three lead compounds and the control is shown in Table 2 and depicted in Fig 8. Residues ILE_10, ALA_32, VAL_18, VAL_64, LEU_83, LEU_134, PHE_80, PHE_82, and ALA_144 were commonly involved in hydrophobic interactions. HIE_84 and GLN_85 frequently participated in polar interactions, while LYS_33 and LYS_89 were consistently involved in hydrogen bonding. These findings indicate that the three lead compounds and the control occupy the same binding pocket. The validation of docking poses of the three lead compounds, CIDs-6474893, 10469828, 135438111, and control CID-44480399 within the binding site of the target protein (CDK2) is displayed in Fig 9.

thumbnail
Download:
Table 2. Comprehensive analysis of intermolecular interactions and key amino acid residues involved in protein-ligand complex formation.

https://doi.org/10.1371/journal.pone.0331438.t002

thumbnail
Download:
Fig 8. Protein-ligand interactions between the CDK2 protein and the three lead compounds, along with the control.

CIDs-6474893 (A), 10469828 (B), 135438111 (C), and 44480399 (control) (D).

https://doi.org/10.1371/journal.pone.0331438.g008

thumbnail
Download:
Fig 9. Validation of docking poses of selected ligands within the binding site of the target protein.

The protein is represented in green cartoon format, with four docked ligands shown in different colors: CID-6474893 (blue, A), CID-10469828 (purple, B), CID-135438111 (cyan, C), and control CID-44480399 (yellow, D). All docked ligands are visualized in complex with the red-colored co-crystallized ligand to confirm accurate binding site alignment.

https://doi.org/10.1371/journal.pone.0331438.g009

3.4 Post-docking MM-GBSA calculation

The MM-GBSA process was used for calculating the binding free energy between the protein and ligand. The target protein CDK-2 interacted with the control medication Fruquintinib and the three lead compounds (CIDs-6474893, 10469828, and 135438111) with higher docking scores. The examination of the MM-GBSA result revealed that they had an increased net negative pattern and higher binding free energy scores. The MM-GBSA results signify that the leads CIDs-6474893, 10469828, 135438111, and the control drug CID-44480399 showed negative binding free energies of −59.79, −55.64, −46.99, and −40.39 kcal/mol, respectively. The lead compounds demonstrated a more stable association with the target protein than the control medication, with CID-6474893 exhibiting the most significant binding energy release (−59.79 kcal/mol), as illustrated in Fig 10. Further study of these three compounds revealed significant activity in ΔG Bind Covalent, ΔG Bind Coulomb, ΔG Bind Lipo, ΔG Bind Hbond, ΔG Bind Packing, ΔG Bind Solv GB, and ΔG Bind vdW (Fig 10 and Supplementary S3).

thumbnail
Download:
Fig 10. The radar chart shows the post-docking MM-GBSA scores of the three lead compounds (CIDs-6474893, 10469828, and 135438111) and the control (CID-44480399).

https://doi.org/10.1371/journal.pone.0331438.g010

3.5 SAR analysis

The chemistry and physical structure of a phytoconstituent significantly influence its biological functions. The physiological activities of medicinal plants typically support their use, which is crucial for evaluating their potential as therapeutic candidates. A SAR analysis was performed utilizing the web-based PASS tool to assess the anticancer effectiveness of the chosen compounds in conjunction with the control fruquintinib, before the MD simulation. The findings of the SAR models are illustrated in Table 3 and Supplementary S4. The three lead compounds possess a biological function associated with anti-cancer activity according to this analysis. The initial lead compound (CID-6474893) exhibited antimutagenic, antineoplastic, and apoptosis agonist properties. The second lead (CID-10469828) exhibited properties such as antineoplastic, antimutagenic, TP53 expression, apoptosis agonist, enhancement, and TNF expression inhibition. Lastly, the third lead compound (CID-135438111) demonstrated TP53 expression enhancement, antineoplastic properties, chemosensitization, radiosensitization, and immunomodulatory activity for the treatment of cancer-associated disorders. These numerous prior biological activities can guarantee that the selected three compounds will undergo a more thorough evaluation. On the other hand, the control ligand (CID-44480399) revealed many properties as well, such as an angiogenesis inhibitor, antineoplastic, vascular endothelial growth factor antagonist, proto-oncogene tyrosine-protein kinase Fgr inhibitor, platelet-derived growth factor receptor kinase inhibitor, and so on.

thumbnail
Download:
Table 3. Predicted biological activities of selected compounds and the control ligand with corresponding Pa and Pi values.

https://doi.org/10.1371/journal.pone.0331438.t003

3.6 FMO analysis

The three lead compounds of CIDs-6474893, 10469828 and 135438111, and the control Fruquintinib CID-44480399 exhibited a HOMO and LUMO energy score of (−0.22474; −0.08597), (−0.22430; −0.08439), (−0.23481; −0.06148), and (−0.23146; −0.07300), respectively in atomic mass units (au) that are showed at Fig 11 and Table 4. The calculated HOMO-LUMO energy gap (ΔE), molecular hardness (η), and molecular softness (S) energy scores of the three leads and control were (3.776, 3.807, 4.716, and 4.312) eV, (1.888, 1.9035, 2.358, and 2.156) eV, and (0.530, 0.525, 0.424, and 0.464) eV-1, respectively. From the analysis of this investigation, it can be stated that CID-6474893 had the highest chemical reactivity, most responsive to electron transfer, and least resistant to electron density change, as evidenced by its lowest HOMO-LUMO energy gap (ΔE), highest softness (S), and lowest hardness (η). On the other hand, CID-135438111 had the largest ΔE, highest η, and lowest S, indicating low chemical reactivity, resistance to change in electron density, and poor responsiveness to interactions. Conversely, CID-10469828 revealed moderate energy scores in ΔE, η, and S, explaining its balanced interaction and chemical reactivity. Comparing the three leads with the control, CID-6474893 and 10469828 are more likely to engage in effective electronic interactions.

thumbnail
Download:
Table 4. HOMO-LUMO energy gap (ΔE), molecular hardness(η), and molecular softness (S) energy score of the three lead compounds CID-6474893, 10469828, 135438111, and the control 44480399.

https://doi.org/10.1371/journal.pone.0331438.t004

thumbnail
Download:
Fig 11. HOMO-LUMO Gap (ΔE), HOMO, and LUMO of the three lead compounds CIDs-6474893. (A), 10469828 (B), 135438111 (C), and control 44480399 (D).

https://doi.org/10.1371/journal.pone.0331438.g011

3.7 MD simulation analysis

The MD simulation was run for 200 ns, utilizing the Desmond package from Schrödinger. The Simulation Interaction Diagram (SID) was then used to examine the RMSD, RMSF, Rg, SASA, protein-ligand interactions, and ligand-protein interactions. Maestro’s simulation event analysis option was utilized to analyze H-bonds. Using the R programming (Bio3D package), PCA and DCCM were determined from the MD trajectories. Besides that, post-simulation MM-GBSA were also performed on the MD trajectories in the initial and terminating stages (0 ns and 200 ns) using the Prime MM-GBSA module of Desmond.

3.7.1 Protein RMSD analysis.

RMSD is frequently applied to establish the equilibration period, quality of biomolecular simulations, and to group conformations that are comparable in MD trajectory analysis [85]. The acceptable range of protein-ligand complex’s average RMSD change is between 1 and 3 Å. If the RMSD value is more than 1–3 Å, there has been a significant conformational change in the protein structure [86]. The RMSD of the three lead compounds and the control has been assessed to identify alterations in protein structure from the initial state. A 200 ns MD simulation was performed to investigate the conformational change of the target macromolecule within the complex of the specified chemicals, including CIDs-6474893, 10469828, 135438111, and 44480399 (control), shown in Table 5, Fig 12A, and Supplementary S5. The highest RMSD values of the apoprotein, selected three lead compounds (CIDs-6474893, 10469828, and 135438111) and control (CID-44480399), were 2.16, 2.27, 2.348, 2.264, and 2.587 Å, respectively, and the lowest scores were 1.187, 0.993, 1.003, 1.095, and 0.934 Å, respectively. The average RMSD scores for apoprotein, three lead compounds (CIDs-6474893, 10469828, and 135438111), and control (CID-44480399) were 1.628, 1.715, 1.683, 1.653, and 1.9 Å, respectively. The average RMSD values of all substances remained within the acceptable range of 1–3 Å. The control (CID-44480399) exhibited excessive fluctuation during the 20–40, 40–80, and 100–120 ns simulation time ranges and failed to maintain a constant fluctuation with the apoprotein throughout the 20–120 ns simulation time range. Compared to the control (CID-44480399), all three lead compounds (CIDs-6474893, 10469828, and 135438111) worked very well when mixed with apoprotein till the conclusion of the simulation session and kept a steady level of fluctuation, as their average RMSD values were less than that of the control.

thumbnail
Download:
Table 5. Three lead compounds, together with the control, produced distinct parameters, such as the greatest, smallest, and average values from the 200 ns MD simulation.

https://doi.org/10.1371/journal.pone.0331438.t005

thumbnail
Download:
Fig 12. Protein and ligand RMSD, RMSF, Rg, SASA, and H-bonds of the ligand-protein complex of CIDs-6474893, 10469828, 135438111, and 44480399 (control), respectively (A, B, C, D, E, F).

https://doi.org/10.1371/journal.pone.0331438.g012

3.7.2 Ligand RMSD analysis.

Ligand RMSD was evaluated to demonstrate the stability of the ligand while associated with the protein binding site. Lower ligand RMSD indicates better stability of a molecule [87]. The ligand RMSD analysis of the three lead compounds (CIDs-6474893, 10469828, and 135438111) and the control (CID-44480399) were shown in Table 5, Fig 12B, and Supplementary S6. According to the results, three lead compounds (CIDs-6474893, 10469828, and 135438111) showed a steadier range of variations, compared to the control, with the lowest and highest values of 0.25 Å in the 42 no. frame to 1.117 Å in the 993 no. frame (the range is 0.867 Å), 0.167 Å in the 260 no. frame to 0.777Å in the 854 no. frame (the range is 0.61 Å), 0.201 Å in the 645 no. frame to 1.156 Å in the 26 no. frame (the range is 0.955 Å), respectively. Consequently, the control displayed a higher fluctuation with a lowest and highest value of 0.269 Å in the 375 no. frame to 1.764 Å in the 781 no. frame (the range is 1.495 Å). The average ligand RMSD scores for three lead compounds (CIDs-6474893, 10469828, and 135438111), along with the control (CID-44480399), were 0.613 Å, 0.432 Å, 0.664 Å, and 1.102 Å, accordingly. All three lead compounds (CIDs-6474893, 10469828, and 135438111) showed lower average ligand RMSD compared to the control (CID-44480399), and among the three lead compounds, CID-10469828 showed the lowest average ligand RMSD, suggesting that CID-10469828 has greater stability than the other compounds (CIDs-6474893 and 135438111). Besides this fact, the difference among the average RMSD values of the three lead compounds was not much, and thus all three of them can be declared as stable compounds compared to the control (CID-44480399).

3.7.3 RMSF analysis.

Utilizing the RMSF approach, the change in protein chain residues and the positioning of ligand atoms can be estimated. The RMSF value of the AA (amino acid) atoms reveals how stable and how much they change in a complicated system. If a compound’s AA residues have low RMSF values, it may be more stable. If a complex’s AA residues have high RMSF values, it may be less stable [86]. The RMSF values of the three lead compounds (CIDs-6474893, 10469828, 135438111) and control (CID-44480399), in complex with the apoprotein, are shown in Table 5, Fig 12C, and Supplementary S7. The greatest RMSF score of apoprotein, three lead compounds (CIDs-6474893, 10469828, and 135438111), and control (CID-44480399) were 4.511, 5.673, 4.516, 6.569, and 3.321 Å, respectively, and the smallest values were 0.337, 0.355, 0.358, 0.364, and 0.362 Å, respectively. The average RMSF values for apoprotein, three lead compounds (CIDs-6474893, 10469828, and 135438111), and control (CID-44480399) were 0.836, 0.881, 0.897, 0.897, and 0.951 Å, respectively. All three lead compounds (CIDs-6474893, 10469828, and 135438111) showed smaller average RMSF compared to the control (CID-44480399). The chosen compounds exhibited a peak area of the protein at THR_14, LEU_25, THR_41, ASH_73, LEU_96, GLU_138, PRO_155, VAL_226, MET_233, and PHE_240 residual sites that showed the greatest fluctuation over the simulation period, as shown in Fig 12C.

3.7.4 Rg analysis.

The Rg may be thought of as the arrangement of atoms along the protein-ligand complex axis. To determine the protein’s stiffness and mobility, the protein-ligand complexes’ Rg were analyzed [86]. A protein molecule with a lower Rg value is said to be closely packed, while one with a larger Rg value is said to be loosely packed [88]. The three lead compounds (CIDs-6474893, 10469828, and 135438111) and the control (CID-44480399) linked to the target protein were screened for stability in terms of Rg throughout the 200 ns simulation period, shown in Table 5, Fig 12D, and Supplementary S6. According to the results, three lead compounds (CIDs-6474893, 10469828, and 135438111) and the control (CID-44480399) showed a range of variations, with the lowest and highest values of 4.422 Å in the 993 no. frame to 5.327 Å in the 305 no. frame (the range is 0.905 Å), 4.434 Å in the 162 no. frame to 4.984 Å in the 309 number frame (the range is 0.55 Å), 2.786 Å in the 820 no. frame to 3.267 Å in the 528 no. frame (the range is 0.481 Å), and 4.803 Å in the 631 no. frame to 5.085 Å in the 141 no. frame (the range is 0.282 Å). The average Rg score for three lead compounds (CIDs-6474893, 10469828, and 135438111), along with the control (CID-44480399), were 4.939, 4.704, 3.123, and 4.963 Å, respectively. According to the result of this investigation, all three lead compounds (CIDs-6474893, 10469828, and 135438111) had lower Rg values compared to the control, but CID-135438111 had the lowest Rg value, indicating that CID-135438111 had the highest stability during the simulation period. These outcomes determine that the protein’s active site fails to undergo significant structural modification when binding to the lead compounds. Meanwhile, the control (CID-44480399) exhibited conformational changes in the protein’s active site. As the lead compounds (CIDs-6474893, 10469828, and 135438111) had a lower Rg value, suggesting that they were tightly packed with the protein, other than the control (CID-44480399). Additionally, CID-135438111, displaying the lowest Rg value, did not fluctuate at all during the simulation time, while the control (CID-44480399) revealed the most notable fluctuation throughout the simulation.

3.7.5 SASA analysis.

The SASA has developed to comprehend the architecture and role of biological macromolecules. On a protein’s surface, the amino acid residues generally serve as active sites and/or interact with other molecules. Analysis of the SASA assists in identifying the solvent-like characteristics (hydrophobic or hydrophilic) of a protein molecule and protein-ligand complexes [86]. A molecule exhibiting a reduced solvent-accessible surface area (SASA) value is tightly compacted with few exposed surfaces, while an elongated or stretched molecule has a greater SASA value [89]. The corresponding SASA values of CIDs-6474893, 10469828, 135438111, and 44480399 (control) were calculated and displayed in Table 5, Fig 12E, and Supplementary S6. According to Fig 12E, three lead compounds (CIDs-6474893, 10469828, and 135438111) showed a steadier range of variations, compared to the control (CID-44480399), with the lowest and highest values of 30.121 Å2 in the 894 no. frame to 126.486 Å2 in the 1 no. frame (the range is 96.365 Å2), 19.969 Å2 in the 682 no. frame to 93.982 Å2 in the 104 no. frame (the range is 74.013 Å2), 5.213 Å2 in the 52 no. frame to 134.051 Å2 in the 744 no. frame (the range is 128.838 Å2), respectively. Consequently, the control displayed a higher fluctuation with the lowest and highest value of 62.859 Å2 in the 1 no. frame to 288.345 Å2 in the 723 no. frame (the range is 225.486 Å2). The three lead compounds (CIDs-6474893, 10469828, and 135438111), along with the control (CID-44480399), had respective average SASA values of 67.673, 51.263, 69.879, and 175.577 Å2. According to this investigation, the reduced average SASA score of the three lead compounds (CIDs-6474893, 10469828, and 135438111) indicates that they are binding in the backbone of the protein, occupying less surface area and more compact binding than the control (CID-44480399). Among the three leads, the CID-10469828 is slightly better, as evidenced by the lowest SASA value, indicating it is more profoundly entrenched inside the protein’s binding pocket in the complex system, hence augmenting its efficacy in regulating the biological activity of the target.

3.7.6 Hydrogen bond analysis.

A drug-binding site that is actively involved in the interaction between the drug and the desired protein and that influences drug adsorption, specificity, and metabolism can be characterized by hydrogen bonds [90]. Therefore, the number of hydrogen bonds of the selected three lead compounds (CIDs-6474893, 10469828, and 135438111), compared with the control (CID-44480399), has been calculated for systems by looking at conformations every 200 ps, as displayed in Table 5, Fig 12F, and Supplementary S6. During the 200 ns simulation period, all compounds concurrently formed numerous hydrogen bonds, with distances varying from 233 to 285. The lowest and highest number of H-bonds of the three lead compounds (CIDs-6474893, 10469828, and 135438111) and the control (CID-44480399) were 241 to 285, 233 to 279, 238 to 279, and 239 to 277, respectively. The average number of H-bonds for the three lead compounds (CIDs-6474893, 10469828, and 135438111) and control (CID-44480399) were 260.676, 255.337, 258.011, and 257.905, respectively. According to the outcome, CID-6474893 and 135438111 had a larger average number of H-bonds than the control (CID-44480399), suggesting that they are more likely to establish hydrogen bonds more forcefully when they bind to the target protein. This could potentially improve drug selectivity.

3.7.7 Protein-ligand interaction analysis.

The SID was used to investigate a protein’s complex structure, assign ligands, and examine interactions between molecules over the 200 ns simulation time. Non-covalent (hydrophobic bonds), ionic, hydrogen bonds, and water bridges were utilized to study and show the protein-ligand complexes. The 200 ns simulation ended with a stable binding with the target protein, as each molecule interacted with it in a different way. CID-6474893 produced more than two contacts at the residues TYR_15, LYS_20, LYS_33, LEU_83, ASP_86, and LYS_89, with interaction fractions (IF) of 0.44, 0.03, 1.02, 1.1, 0.65, and 0.85, respectively (Fig 13A). In Figs 13 and 14, the simulation yields a robust specific interaction as the identical subtype engages with the ligand repeatedly. The residues TYR_15 (0.47), LYS_33 (1.1), LEU_83 (1.0), and ASP_86 (0.9) produced multiple contacts with CID-10469828 (Fig 13B). CID-135438111 also made multiple interactions at residues ILE_10 (0.5) and ASP_86 (0.75) (Fig 13C). The control (CID-44480399) produced various contacts at the residues, LYS_20 (0.25), LYS_33 (0.65), LEU_83 (0.85), LYS_89 (0.53) (Fig 13D). Among the selected compounds (CIDs-6474893, 10469828, and 135438111) and the control (CID-44480399), the desired lead compound (CID-6474893) produced multiple contacts at the residues with a higher interaction fraction, suggesting that it is showing promising interaction with the target protein.

thumbnail
Download:
Fig 13. The protein–ligand interface’s diverse bonding patterns are demonstrated during the 200 ns MD simulation.

The three lead compounds, CIDs-6474893 (A), 10469828 (B), and 135438111 (C), as well as the control, CID-44480399 (D), are displayed.

https://doi.org/10.1371/journal.pone.0331438.g013

thumbnail
Download:
Fig 14. 2D interactions of the ligand–protein complex after a 200 ns simulation.

The three lead compounds and control: CIDs-6474893 (A), 10469828 (B), 135438111 (C), and 44480399 (D) are presented here.

https://doi.org/10.1371/journal.pone.0331438.g014

3.7.8 Ligand-protein interaction analysis.

All three lead compounds (CIDs-6474893, 10469828, and 135438111) and the control (CID-44480399) have been shown to interact with one another in the protein space of the SID. When compared to alternative lead compounds (CIDs-10469828 and 135438111), the desired compound (CID-6474893) exhibited the highest stability due to its strong hydrogen bonding network, hydrophobic interactions, and engagement with key amino acid residues during the simulation. Accordingly, repeated interactions with the same kind of ligand maintain the distinct interaction (Fig 14). In comparison to the control (CID-44480399), compound CID-6474893 showed improved stability in the ligand-protein interaction investigation.

3.7.9 Post-simulation MM-GBSA Analysis.

In this investigation, the MM-GBSA approach was used to determine the target protein ligand-binding free energy, as shown in Fig 15. In order to determine the MM-GBSA of the protein-ligand complex structure, the 200-ns dynamic simulation trajectory was used. The negative binding free energy (ΔG Bind) values of the three lead compounds (CIDs-6474893, 10469828, and 135438111) and the control (CID-44480399) in complex with the targeted protein were (−34.43 and −30.66), (−45.16 and −71.11), (−44.91 and −42.70), and (−61.79 and −24.94) kcal/mol, respectively, at the beginning and end of the experiment (200 ns) (Fig 15 and Supplementary S8). These ΔG Bind values represent the evolution of their stability during molecular dynamics (MD) simulations. The control exhibited greater stability at the outset, but it experienced a substantial decline in stability over time, indicating a weakened binding in the long term. As a result, CID-6474893 maintained a relatively stable state, with only a minor increase in free energy, but a lower stability reduction than the control. Conversely, CID-135438111 maintained a virtually identical level of stability, with only a minor increase in free energy. However, CID-10469828 demonstrated a substantial increase in stability, with a significant decrease in free energy during the simulation, suggesting that it has a superior binding affinity compared to the control and other lead compounds.

thumbnail
Download:
Fig 15. Post-simulation MM-GBSA analysis of the three lead compounds and the control: CIDs-6474893 (A), 10469828 (B), 135438111 (C), and 44480399 (D).

https://doi.org/10.1371/journal.pone.0331438.g015

3.7.10 PCA.

PCA was conducted on the trajectory data of the three lead compounds CIDs-6474893, 10469828, 135438111, the control 44480399, and the apoprotein (PDB ID-6GUE) to examine the protein-ligand complexes’ conformational dynamics during MD simulations. The dynamics and structures of the complexes were identified by analysing the first three principal components (PCs) [88]. CID-6474893 (Fig 16A) had 21.72% variance from PC1, 9.61% from PC2, and 9.32% from PC3, totaling 40.65%. A significant categorization of conformations along these PCs shows considerable conformational differences across states. The eigenvalue scree plot shows a progressive decrease in variance contribution, implying a dispersed component motion contribution. In comparison, for CID-10469828 (Fig 16B), PC1 contributed 21.78% of the variance, followed by PC2 (11.52%) and PC3 (6.94%), accounting for a total variance of 40.24%. Conversely, CID-135438111 (Fig 16C) displayed a variance of 24.67% in PC1, 9.52% in PC2, and 6.68% in PC3, totaling 40.87%. In contrast, CID-44480399 (control) (Fig 16D) exhibited a cumulative variance of approximately 36.58%, equivalent to 21.97% from PC1, 8.76% from PC2, and 5.85% from PC3. Finally, the apoprotein (PDB ID-6GUE) (Fig 16E) demonstrated a variance of 13.75% in PC1, followed by 6.8% in PC2, and 6.34% in PC3, which collectively accounted for 26.89% of the total variance. The substantial conformational stability and a well-defined binding mode are suggested by the highly distinct separation of clusters along PC1 and PC2. The eigenvalue scree plot indicates that the variance contribution decreases rapidly beyond the first three PCs, indicating that the majority of conformational motions are captured within these components. In comparison to the apoprotein, all three lead compounds, CIDs-6474893, 10469828, 135438111, and the control 44480399, showed a higher proportion of variance in the first few PCs when complexed with the protein. Among them, CID-135438111 showed the least flexible and most stable binding conformation, as evidenced by a significant amount of variation. This suggests that CID-135438111 is the most stable among the analyzed complexes, followed by CID-6474893, 10469828, and the control 44480399, which exhibited relatively higher conformational fluctuations.

thumbnail
Download:
Fig 16. Comparing the eigenvalue using PCA and the variance percentage.

Each area is shown on one of three panels. PC1, PC2, and PC3 are three variants. At this location, CIDs-6474893 (A), 10469828 (B), 135438111 (C), control 44480399 (D), and apoprotein (PDB ID-6GUE) (E).

https://doi.org/10.1371/journal.pone.0331438.g016

3.7.11 DCCM analysis.

Protein motions occur on various time scales, including femtoseconds and seconds. DCCM is also affected by the time at which the correlation data was gathered [91]. Using inter-residue DCCM analysis, the target protein (PDB ID-6GUE), together with its docked complexes CIDs-6474893, 10469828, 135438111, control 44480399, and apoprotein (PDB ID-6GUE), were examined for correlated and anti-correlated moves. The colour blue denotes residues that are related, while the colour sea green denotes residues that are anti-correlated. The residue index maps indicated that the target protein (PDB ID-6GUE) was dispersed across populations through both positive and negative correlations. A cross-correlation map was generated by comparing the pairwise cross-correlation coefficients (Fig 17) between the target protein (PDB ID-6GUE) and the three lead compounds CIDs-6474893, 10469828, 135438111, and the control 44480399. The analysis revealed that CID-10469828 and 44480399 (control) had more positively correlated regions than CID-6474893 and 135438111. However, those compounds (CID-10469828 and 44480399) also exhibited a higher number of negatively correlated regions in comparison to the other two lead compounds (CIDs-6474893 and 135438111). In contrast, CID-135438111 exhibited a higher number of positively correlated regions between these two lead compounds (CID-6474893 and 135438111), while CID-6474893 had a lower number of negatively correlated regions. As a consequence, CID-6474893 demonstrated greater stability than the control (CID-44480399) and the other two lead compounds (CID-10469828 and 135438111).

thumbnail
Download:
Fig 17. DCCM is represented by the colours blue and sea green, which signify positive and negative, respectively.

Here, CIDs-6474893 (A), 10469828 (B), 135438111 (C), 44480399 (control) (D), apoprotein (PDB ID-6GUE) (E).

https://doi.org/10.1371/journal.pone.0331438.g017

4. Discussion

The alarmingly high incidence of colorectal cancer nowadays is quickly turning into a major issue in public health [92]. It has been difficult to create new pharmacologically effective anti-cancer treatment agents due to the lack of molecular knowledge about the oncogenic initiation and the interplay of biochemical processes and components. Recurrence is still a possibility with standard chemotherapy approaches, despite their remarkable efficacy in treating CRC [93].

Natural products are a great place to find new medicinal leads, as they include a wide variety of chemical components [94]. Novel drugs, especially those used to treat malignant illnesses, owe a great deal to their contributions [95]. Recent advancements in CRC treatment include breakthroughs in phytochemical formulations [96,97]. According to the studies, EGCG (epigallocatechin gallate) found in green tea, reduces tumor growth by inhibiting VEGFR-2 and modulating Wnt/β-catenin pathways in animal models. It also synergizes with chemotherapy to suppress cancer stem cells. Besides that, curcumin has also been shown to exhibit effects such as pro-apoptotic and anti-inflammatory via NF-κB inhibition [96]. Nanoformulations enhance their bioavailability and efficacy in CRC xenografts [97].

This study used phytochemicals from the IMPPAT database that were antimutagenic, antineoplastic, and apoptotic agonists. The effectiveness of these plant components in colon cancer therapy is still uncertain. Computational approaches were used in this investigation to find cyclin-dependent kinase (CDK) inhibitors that prevent malignant cell development and colorectal cancer. HTVS and MD simulations are effective and fast bioinformatics tools for screening multiple drug candidates [87,98]. In the present research, the goal of identifying the CDK-2 (2VSM) was to find a potential therapeutic candidate. Additionally, out of 4433 drug-like compounds, five (CIDs-135438111, 6474893, 44257567, 10469828, and 353825) exhibited increased binding affinities compared to the control (CID-44480399). These compounds were identified by protein production, ligand creation, and molecular docking studies. Furthermore, these high-affinity ligands were advanced through subsequent ADME/T profiling to assess their efficacy as potential lead compounds. Of these, three phytochemicals (CIDs-6474893, 10469828, and 135438111) demonstrated excellent properties. They did not pose any toxicological risks, and exhibited satisfactory pharmacokinetics and Lipinski criteria maintenance. Additionally, the MWs (molecular weights) of these compounds were lower compared to the control ligand. Heat molecules benefited from this since compounds with greater MWs violate the Lipinski rule [99]. The logP value is a determinant of the drug molecule’s assimilation into the organism [100]. The three compounds that were selected demonstrated logP values within the admissible range, as determined by the Lipinski criterion. The docking results were also validated using the post-docking MM-GBSA analysis. This investigation revealed that the selected three leads had more negative binding free energies than the control, which explains the greater binding affinity and the conformational stability of the leads. After that, the HOMO and LUMO energy scores were calculated for the selected compounds before the MD simulation. According to this analysis, CID-6474893 fits as the best candidate due to its highest chemical reactivity, most responsiveness to electron transfer, and least resistance to electron density change. Nevertheless, these quantum chemical descriptors obtained from this FMO analysis do not fully represent biological complexities such as solvent effects and protein flexibility. Consequently, a lower ΔE suggests the possibility of reactivity; however, it should be regarded as a supplementary factor to biological efficacy rather than a predictor of it. In this investigation, docking, MM-GBSA, ADME/T, SAR, and MD simulations were integrated with HOMO–LUMO results to facilitate a thorough assessment. In order to verify these discoveries, further experimental validation is required. For that reason, if these compounds show promise as a CRC treatment option based on MD modeling results for binding stability with the target protein, they could be worth considering. However, before starting clinical trials, it is crucial to perform in vivo and in vitro investigations. Using MD simulation, one can determine the stability and stiffness of ligand-protein complexes in an environment other than the human body. A 200 ns simulation was performed on the chosen compounds to determine the stability of their ligand-protein complexes. Using MD simulation, the following parameters of the control ligand, three lead compounds, and the protein-ligand interaction were evaluated: RMSD, RMSF, Rg, SASA, H-bonds, PCA, and DCCM. Post-simulation MM-GBSA was also run. All three lead compounds had very stable RMSDs when compared to the reference. The RMSF results showed that none of the three compounds were as unstable as the control. Based on the analysis, the average Rg of the leads was lower than the control ligand. Among them, the Rg value of CID-135438111 was the lowest, indicating that CID-135438111 can tightly pack with the protein, other than CID-44480399 (control). In the meantime, the three lead compounds had reduced SASA values compared to the control, and among the three of them, CID-10469828 displayed the lowest SASA value, indicating that it is more deeply embedded in the protein’s binding pocket in the complex system, thereby increasing its effectiveness in modulating the biological activity of the target. In the H-bond analysis, the two lead compounds (CID-6474893 and 135438111) were likely to form hydrogen bonds more strongly while binding with the target protein than the control (CID-44480399), due to their higher average H-bond values. Furthermore, in protein-ligand and ligand-protein contact analysis, CID-6474893 showed the highest stability due to its strong hydrogen bonding network, hydrophobic interactions, and engagement with key amino acid residues during the simulation. The protein-ligand binding free energy was estimated using the 200 ns dynamic simulation trajectory at both the initial and final stages (0 ns to 200 ns). According to this analysis, the control showed higher stability in the initial stage of the simulation, but it lost its stability over time, suggesting weaker binding in the long run. Comparatively, the lead compound CID-10469828 showed the best binding stability, as its binding free energy became more negative over time, suggesting enhanced stability after 200 ns. The other two lead compounds (CID-6474893 and 135438111) showed stable binding, compared to the control, but they were less favorable than CID-10469828. According to the PCA analysis, CID-135438111 showed the least flexible and most stable binding conformation, followed by CID-6474893, 10469828, and the control 44480399, which exhibited relatively higher conformational fluctuations. The DCCM study demonstrated that CID-6474893 had a higher level of stability than the control (CID-44480399), as evidenced by the fact that it had a lower number of negatively correlated regions. Despite the fact that CID-44480399 (control) had a greater number of positively correlated regions than the three lead compounds, it also had a greater number of negatively correlated regions, which exceeded the criteria of having a higher number of positively correlated regions. Thus, it can be concluded that the control did not perform well in any of the analyses in comparison to the three lead compounds (CIDs-6474893, 10469828, and 135438111), and CID-6474893 performed best among them. Here, CID-6474893 ((1E,4E)-1,5-bis(4-hydroxy-3-methoxyphenyl) penta-1,4-dien-3-one) and CID-10469828 ((1E,4E)-1-(4-Hydroxy-3-methoxyphenyl)-5-(4-hydroxyphenyl)-1,4-pentadien-3-one) were retrieved from Curcuma longa, whether CID-135438111 was from Peganum harmala. As these lead compounds have not been previously reported, their identification in this study marks the first instance of their recruitment, suggesting their potential for further research in the development of CRC treatment.

In cheminformatics-based drug design, we used a computer-aided algorithm to simulate the biological system. But in reality, the biological system is quite complex, and the dynamic nature of protein flexibility is unstable [101]. An in silico model can’t count all off-target effects as it doesn’t deal with the whole cellular system. In a biological system, the binding of a ligand to a receptor initiates a downstream signaling cascade, whereas in silico analysis focuses solely on its targeted effect or process. ADME/T analysis is a prediction-based method to calculate all its possible outcomes. But in reality, it doesn’t reflect a drug’s pharmacokinetic properties. It can only give a primary guess. So, in vitro and in vivo validation is a crying need for discovering a new therapeutic drug. An in vitro model works like a bridge between computational predictions and real-world biology, and in vivo analysis shows the overall effect of a drug in a living organism, where all physiological processes of a biological system are interconnected [102]. Our in silico drug design method against CRC treatment was completely a computer-based analysis, which is also known as a pre-clinical study. As a consequence of our laboratory support and funding, we were unable to expand our research findings. Consequently, individuals who are involved in the treatment of this disease may utilize this outcome as a basis for their in vivo/in vitro validation.

5. Conclusion

Cheminformatics-based drug design was employed to simulate the druggability of phytochemicals against colorectal cancer. Phytochemicals that target the enzyme CDK-2 may prevent the progression of cancer through halting the spread of tumors and killing off cancer-causing stem cells by influencing their activity. The main objective of this study is to identify potential lead chemicals that can inhibit cancer progression. Because of this, the objective of the investigation is to determine whether the leads exhibit superior efficacy in preventing the progression of cancer compared to a commercially available drug. In this investigation, molecular docking, ADME/T, protein-ligand interaction, MM-GBSA, SAR, FMO, and MD simulations identified three compounds (CIDs-6474893, 10469828, and 135438111) that have the potential to inhibit the activity of CDK-2, thereby facilitating the development of an anticancer drug. However, in order to verify the compounds’ efficacy against cancer, in vivo and in vitro studies are necessary.

Supporting information

S1 File. Anticancer plant list.

https://doi.org/10.1371/journal.pone.0331438.s001

(ZIP)

Acknowledgments

We are grateful to the Bioinformatics Laboratory (BioLab), Bangladesh for providing software support, training, and assistance during the work.

References

  1. 1. Siegel RL, Giaquinto AN, Jemal A. Cancer statistics, 2024. CA Cancer J Clin. 2024;74(1):12–49.
  2. 2. Weitz J, Koch M, Debus J, Höhler T, Galle PR, Büchler MW. Colorectal cancer. Lancet. 2005;365(9454):153–65. pmid:15639298
  3. 3. Siegel RL, Miller KD, Goding Sauer A, Fedewa SA, Butterly LF, Anderson JC. Colorectal cancer statistics, 2020. CA Cancer J Clin. 2020;70(3):145–64.
  4. 4. Ladabaum U, Mannalithara A, Meester RGS, Gupta S, Schoen RE. Cost-Effectiveness and National Effects of Initiating Colorectal Cancer Screening for Average-Risk Persons at Age 45 Years Instead of 50 Years. Gastroenterology. 2019;157(1):137–48. pmid:30930021
  5. 5. Ferlay J, Soerjomataram I, Dikshit R, Eser S, Mathers C, Rebelo M, et al. Cancer incidence and mortality worldwide: sources, methods and major patterns in GLOBOCAN 2012. Int J Cancer. 2015;136(5):E359-86. pmid:25220842
  6. 6. Simon L, Imane A, Srinivasan KK, Pathak L, Daoud I. In Silico Drug-Designing Studies on Flavanoids as Anticolon Cancer Agents: Pharmacophore Mapping, Molecular Docking, and Monte Carlo Method-Based QSAR Modeling. Interdiscip Sci. 2017;9(3):445–58. pmid:27059855
  7. 7. Li J, Ma X, Chakravarti D, Shalapour S, DePinho RA. Genetic and biological hallmarks of colorectal cancer. Genes Dev. 2021;35(11–12):787–820. pmid:34074695
  8. 8. Murphy N, Moreno V, Hughes DJ, Vodicka L, Vodicka P, Aglago EK, et al. Lifestyle and dietary environmental factors in colorectal cancer susceptibility. Mol Aspects Med. 2019;69:2–9. pmid:31233770
  9. 9. Sawicki T, Ruszkowska M, Danielewicz A, Niedźwiedzka E, Arłukowicz T, Przybyłowicz KE. A Review of Colorectal Cancer in Terms of Epidemiology, Risk Factors, Development, Symptoms and Diagnosis. Cancers (Basel). 2021;13(9):2025. pmid:33922197
  10. 10. Kolligs FT. Diagnostics and Epidemiology of Colorectal Cancer. Visc Med. 2016;32(3):158–64. pmid:27493942
  11. 11. Olde Bekkink M, McCowan C, Falk GA, Teljeur C, Van de Laar FA, Fahey T. Diagnostic accuracy systematic review of rectal bleeding in combination with other symptoms, signs and tests in relation to colorectal cancer. Br J Cancer. 2010;102(1):48–58. pmid:19935790
  12. 12. Xie Y-H, Chen Y-X, Fang J-Y. Comprehensive review of targeted therapy for colorectal cancer. Signal Transduct Target Ther. 2020;5(1):22. pmid:32296018
  13. 13. Zaman A, Khan MSS, Akter L, Syeed SH, Akter J, Al Mamun A, et al. Exploring new pharmacology and toxicological screening and safety evaluation of one widely used formulation of Nidrakar Bati from South Asia region. BMC Complement Altern Med. 2015;15:121. pmid:25880852
  14. 14. Afrin S, Giampieri F, Gasparrini M, Forbes-Hernández TY, Cianciosi D, Reboredo-Rodriguez P, et al. Dietary phytochemicals in colorectal cancer prevention and treatment: A focus on the molecular mechanisms involved. Biotechnol Adv. 2020;38:107322. pmid:30476540
  15. 15. Yang L, Atakhanova N, Arellano MTC, Mohamed MY, Hani T, Fahdil AA, et al. Translational research of new developments in targeted therapy of colorectal cancer. Pathol Res Pract. 2023;252:154888. pmid:37948996
  16. 16. Li J, Qin S, Xu R-H, Shen L, Xu J, Bai Y, et al. Effect of Fruquintinib vs Placebo on Overall Survival in Patients With Previously Treated Metastatic Colorectal Cancer: The FRESCO Randomized Clinical Trial. JAMA. 2018;319(24):2486–96. pmid:29946728
  17. 17. Liu S, Lu L, Pan F, Yang C, Liang J, Liu J, et al. Real-World Data: Fruquintinib in Treating Metastatic Colorectal Cancer. Oncol Res. 2022;29(1):25–31. pmid:35063062
  18. 18. Stucchi E, Bartolini M, Airoldi M, Fazio R, Daprà V, Mondello G, et al. Fruquintinib as new treatment option in metastatic colorectal cancer patients: is there an optimal sequence?. Expert Opin Pharmacother. 2024;25(4):371–82. pmid:38568032
  19. 19. Deng Y, Li X. Fruquintinib and its use in the treatment of metastatic colorectal cancer. Future Oncol. 2019;15(22):2571–6. pmid:31407939
  20. 20. Zhao S, Wang W, Li J, Li Z, Liu Z, Zhang S. Clinical research progress of fruquintinib in the treatment of malignant tumors. Invest New Drugs. 2024.
  21. 21. Li Q, Cheng X, Zhou C, Tang Y, Li F, Zhang B, et al. Fruquintinib Enhances the Antitumor Immune Responses of Anti-Programmed Death Receptor-1 in Colorectal Cancer. Front Oncol. 2022;12:841977. pmid:35371995
  22. 22. Guo Y, Zhang W, Ying J, Zhang Y, Pan Y, Qiu W, et al. Phase 1b/2 trial of fruquintinib plus sintilimab in treating advanced solid tumours: The dose-escalation and metastatic colorectal cancer cohort in the dose-expansion phases. Eur J Cancer. 2023;181:26–37. pmid:36628898
  23. 23. Naeem A, Hu P, Yang M, Zhang J, Liu Y, Zhu W. Natural Products as Anticancer Agents: Current Status and Future Perspectives. Molecules. 2022;27(23).
  24. 24. Dehelean CA, Marcovici I, Soica C, Mioc M, Coricovac D, Iurciuc S, et al. Plant-Derived Anticancer Compounds as New Perspectives in Drug Discovery and Alternative Therapy. Molecules. 2021;26(4).
  25. 25. Fang J, Liu C, Wang Q, Lin P, Cheng F. In silico polypharmacology of natural products. Brief Bioinform. 2017;19(6):1153–71.
  26. 26. Chinthakunta N, Cheemanapalli S, Chinthakunta S, Anuradha CM, Chitta SK. A new insight into identification of in silico analysis of natural compounds targeting GPR120. Netw Model Anal Health Inform Bioinform. 2018;7(1):8. pmid:29780684
  27. 27. Macharia JM, Káposztás Z, Bence RL. Medicinal Characteristics of Withania somnifera L. in Colorectal Cancer Management. Pharmaceuticals (Basel). 2023;16(7):915. pmid:37513827
  28. 28. Singh AK, Sharma N, Ghosh M, Park YH, Jeong DK. Emerging importance of dietary phytochemicals in fight against cancer: Role in targeting cancer stem cells. Critical Reviews in Food Science and Nutrition. 2017;57:3449–63.
  29. 29. Islam MR, Awal MA, Khames A, Abourehab MAS, Samad A, Hassan WMI. Computational identification of druggable bioactive compounds from Catharanthus roseus and Avicennia marina against colorectal cancer by targeting thymidylate synthase. Molecules. 2022;27(7).
  30. 30. Illian DN, Hafiz I, Meila O, Utomo ARH, Nuryawan A, Siregar GA. Current status, distribution, and future directions of natural products against colorectal cancer in Indonesia: A systematic review. Molecules. 2021;26(16):1–18.
  31. 31. Alhajjat AM, Lee AE, Strong BS, Shaaban AF. NK cell tolerance as the final endorsement of prenatal tolerance after in utero hematopoietic cellular transplantation. Front Pharmacol. 2015;6:51. pmid:25852555
  32. 32. Islam MR, Akash S, Rahman MM, Nowrin FT, Akter T, Shohag S, et al. Colon cancer and colorectal cancer: Prevention and treatment by potential natural products. Chem Biol Interact. 2022;368:110170. pmid:36202214
  33. 33. K Aziz B, M Salh D, Kaufhold S, Bertier P. The High Efficiency of Anionic Dye Removal Using Ce-Al13/Pillared Clay from Darbandikhan Natural Clay. Molecules. 2019;24(15):2720. pmid:31357459
  34. 34. Christensen JB. Bach Goes to Town. Biomolecules. 2018;8(3):75. pmid:30127319
  35. 35. Redondo-Blanco S, Fernández J, Gutiérrez-Del-Río I, Villar CJ, Lombó F. New Insights toward Colorectal Cancer Chemotherapy Using Natural Bioactive Compounds. Front Pharmacol. 2017;8:109. pmid:28352231
  36. 36. Shi X-N, Li H, Yao H, Liu X, Li L, Leung K-S, et al. Adapalene inhibits the activity of cyclin-dependent kinase 2 in colorectal carcinoma. Mol Med Rep. 2015;12(5):6501–8. pmid:26398439
  37. 37. Tadesse S, Anshabo AT, Portman N, Lim E, Tilley W, Caldon CE, et al. Targeting CDK2 in cancer: challenges and opportunities for therapy. Drug Discov Today. 2020;25(2):406–13. pmid:31839441
  38. 38. García-Reyes B, Kretz A-L, Ruff J-P, von Karstedt S, Hillenbrand A, Knippschild U, et al. The Emerging Role of Cyclin-Dependent Kinases (CDKs) in Pancreatic Ductal Adenocarcinoma. Int J Mol Sci. 2018;19(10):3219. pmid:30340359
  39. 39. Du W, Searle J. The Rb Pathway and Cancer Therapeutics. Curr Drug Targets. 2012;10(7):581–9.
  40. 40. Akiyama T, Ohuchi T, Sumida S, Matsumoto K, Toyoshima K. Phosphorylation of the retinoblastoma protein by cdk2. Proc Natl Acad Sci U S A. 1992;89(17):7900–4. pmid:1518810
  41. 41. Egger JV, Lane MV, Antonucci LA, Dedi B, Krucher NA. Dephosphorylation of the Retinoblastoma protein (Rb) inhibits cancer cell EMT via Zeb. Cancer Biol Ther. 2016;17(11):1197–205. pmid:27645778
  42. 42. Wood DJ, Korolchuk S, Tatum NJ, Wang L-Z, Endicott JA, Noble MEM, et al. Differences in the Conformational Energy Landscape of CDK1 and CDK2 Suggest a Mechanism for Achieving Selective CDK Inhibition. Cell Chem Biol. 2019;26(1):121-130.e5. pmid:30472117
  43. 43. Chagaleti BK, B SK, GV A, Rajagopal R, Alfarhan A, Arockiaraj J, et al. Targeting cyclin-dependent kinase 2 CDK2: Insights from molecular docking and dynamics simulation – A systematic computational approach to discover novel cancer therapeutics. Comput Biol Chem [Internet]. 2024;112:108134.
  44. 44. Oh SJ, Erb HHH, Hobisch A, Santer FR, Culig Z. Sorafenib decreases proliferation and induces apoptosis of prostate cancer cells by inhibition of the androgen receptor and Akt signaling pathways. Endocr Relat Cancer. 2012;19(3):305–19. pmid:22383427
  45. 45. Dachineni R, Ai G, Kumar DR, Sadhu SS, Tummala H, Bhat GJ. Cyclin A2 and CDK2 as Novel Targets of Aspirin and Salicylic Acid: A Potential Role in Cancer Prevention. Mol Cancer Res. 2016;14(3):241–52. pmid:26685215
  46. 46. Tadesse S, Caldon EC, Tilley W, Wang S. Cyclin-Dependent Kinase 2 Inhibitors in Cancer Therapy: An Update. J Med Chem. 2019;62(9):4233–51. pmid:30543440
  47. 47. Chen Z, Jiang L. The clinical application of fruquintinib on colorectal cancer. Expert Rev Clin Pharmacol. 2019;12(8):713–21. pmid:31177854
  48. 48. Weidenhammer E, Tsongalis GJ. An Overview of Molecular Genetics. Mol Diagnostics. 2006;4:259–66.
  49. 49. Daina A, Michielin O, Zoete V. SwissADME: a free web tool to evaluate pharmacokinetics, drug-likeness and medicinal chemistry friendliness of small molecules. Sci Rep. 2017;7:42717. pmid:28256516
  50. 50. Banerjee P, Eckert AO, Schrey AK, Preissner R. ProTox-II: a webserver for the prediction of toxicity of chemicals. Nucleic Acids Res. 2018;46(W1):W257–63. pmid:29718510
  51. 51. Kumar BK, Venkata K, Chandra G, Ojha R, Prajapati VK, Pai A. Pharmacophore based virtual screening, molecular docking, molecular dynamics and MM- GBSA approach for identification of prospective SARS-CoV-2 inhibitor from natural product databases. J Biomol Struct Dyn. 2020;0(0):1–24.
  52. 52. Wang E, Sun H, Wang J, Wang Z, Liu H, Zhang JZH, et al. End-Point Binding Free Energy Calculation with MM/PBSA and MM/GBSA: Strategies and Applications in Drug Design. Chem Rev. 2019;119(16):9478–508. pmid:31244000
  53. 53. Gilson MK, Given JA, Bush BL, McCammon JA. The statistical-thermodynamic basis for computation of binding affinities: a critical review. Biophys J. 1997;72(3):1047–69. pmid:9138555
  54. 54. Islam MR, Tayyeb JZ, Paul HK, Islam MN, Oduselu GO, Bayıl I, et al. In silico analysis of potential inhibitors for breast cancer targeting 17beta-hydroxysteroid dehydrogenase type 1 (17beta-HSD1) catalyses. J Cell Mol Med. 2024;28(15):e18584. pmid:39135338
  55. 55. Anowar Hosen M, Alam A, Islam M, Fujii Y, Ozeki Y, Abe Kawsar SM. Geometrical optimization, PASS prediction, molecular docking, and in silico ADMET studies of thymidine derivatives against FimH adhesin of Escherichia coli. Bulg Chem Commun. 2021;53(3):327–42.
  56. 56. Khalid M, Tayyab F, Adeel M, Tahir N, Braga AAC, Alrashidi KA. Exploration of selenophene analogue and different acceptor influence on photovoltaic properties of pyrrole-4,6(5-H)-dione based chromophores via quantum chemical investigations. Scientific Reports. 2025;15(1):14792.
  57. 57. Hussain R, Khan MU, Mehboob MY, Khalid M, Iqbal J, Ayub K, et al. Enhancement in Photovoltaic Properties of N,N‐diethylaniline based Donor Materials by Bridging Core Modifications for Efficient Solar Cells. ChemistrySelect. 2020;5(17):5022–34.
  58. 58. Zhuo L, Liao W, Yu Z. A Frontier Molecular Orbital Theory Approach to Understanding the Mayr Equation and to Quantifying Nucleophilicity and Electrophilicity by Using HOMO and LUMO Energies. Asian J Org Chem. 2012;1(4):336–45.
  59. 59. Padmaja L, Ravikumar C, Sajan D, Hubert Joe I, Jayakumar VS, Pettit GR, et al. Density functional study on the structural conformations and intramolecular charge transfer from the vibrational spectra of the anticancer drug combretastatin‐A2. J Raman Spectroscopy. 2008;40(4):419–28.
  60. 60. Frisch MJ, Trucks GW, Schlegel HB, Scuseria GE, Robb MA, Cheeseman JR, et al. Gaussian 09, Revision D.01, Gaussian. Inc, Wallingford CT [Internet]. 2009;201. https://cir.nii.ac.jp/crid/1370298757422456580
    • 61. Dennington RDII, Keith TA, Millam JM. GaussView, Version 6.0.16. Shawnee Mission, KS. 2016.
      • 62. Iqbal MMA, Hussain M, Hussain R, Ashraf W. Designing of pyrrolopyrazine-based electron transporting materials with architecture (A1-D-A2) in perovskite solar cells: a DFT study. Scientific Reports. 2025;15(1):15403.
      • 63. Huwaimel B, Alqarni S. A DFT investigation of the adsorption mechanism of paclitaxel on functionalized graphene oxide for enhanced drug delivery. Sci Rep. 2025;15(1):13889. pmid:40263566
      • 64. Finley JP. Using the local density approximation and the LYP, BLYP and B3LYP functionals within reference-state one-particle density-matrix theory. Molecular Physics. 2004;102(7):627–39.
      • 65. Choudhary V, Bhatt A, Dash D, Sharma N. DFT calculations on molecular structures, HOMO-LUMO study, reactivity descriptors and spectral analyses of newly synthesized diorganotin(IV) 2-chloridophenylacetohydroxamate complexes. J Comput Chem. 2019;40(27):2354–63. pmid:31246313
      • 66. Parr RG, Pearson RG. Absolute hardness: companion parameter to absolute electronegativity. J Am Chem Soc. 1983;105(26):7512–6.
      • 67. Koopmans T. Über die Zuordnung von Wellenfunktionen und Eigenwerten zu den Einzelnen Elektronen Eines Atoms. Physica. 1934;1(1–6):104–13.
      • 68. Sagir M, Mushtaq K, Khalid M, Khan M, Tahir MB, Braga AAC. Exploration of linear and third-order nonlinear optical properties for donor-π-linker-acceptor chromophores derived from ATT-2 based non-fullerene molecule. RSC Adv. 2023;13(45):31855–72. pmid:37920195
      • 69. Soleimani Amiri S, Makarem S, Ahmar H, Ashenagar S. Theoretical studies and spectroscopic characterization of novel 4-methyl-5-((5-phenyl-1,3,4-oxadiazol-2-yl)thio)benzene-1,2-diol. Journal of Molecular Structure. 2016;1119:18–24.
      • 70. Malek A, Balawender R. Revisiting the chemical reactivity indices as the state function derivatives. The role of classical chemical hardness. J Chem Phys. 2015;142(5):054104. pmid:25662633
      • 71. Shahab S, Sheikhi M, Filippovich L, Khaleghian M, Dikusar E, Yahyaei H. Spectroscopic studies (geometry optimization, E → Z isomerization, UV/Vis, excited states, FT-IR, HOMO-LUMO, FMO, MEP, NBO, polarization) and anisotropy of thermal and electrical conductivity of new azomethine dyes in stretched polymer matrix. Silicon. 2018;10(5):2361–85.
      • 72. Srivastava HK, Pasha FA, Singh PP. Atomic softness‐based QSAR study of testosterone. Int J of Quantum Chemistry. 2005;103(3):237–45.
      • 73. Shiroudi A, Śmiechowski M, Czub J, Abdel-Rahman MA. Computational analysis of substituent effects on proton affinity and gas-phase basicity of TEMPO derivatives and their hydrogen bonding interactions with water molecules. Sci Rep. 2024;14(1):8434. pmid:38600208
      • 74. De Vivo M, Masetti M, Bottegoni G, Cavalli A. Role of Molecular Dynamics and Related Methods in Drug Discovery. J Med Chem. 2016;59(9):4035–61. pmid:26807648
      • 75. Bai Q, Liu S, Tian Y, Xu T, Banegas-Luna AJ, Pérez-Sánchez H. Application advances of deep learning methods for de novo drug design and molecular dynamics simulation. Wiley Interdiscip Rev Comput Mol Sci. 2022;12(3):1–19.
      • 76. Bowers KJ, Chow E, Xu H, Dror RO, Eastwood MP, Gregersen BA, et al. Scalable algorithms for molecular dynamics simulations on commodity clusters. In: Proceedings of the 2006 ACM/IEEE Conference on Supercomputing. New York, NY, USA: Association for Computing Machinery; 2006. p. 84–es. (SC ’06).
        • 77. Jalali A, Dabaghian F, Zarshenas MM. Alkaloids of Peganum harmala: Anticancer Biomarkers with Promising Outcomes. Curr Pharm Des. 2021;27(2):185–96. pmid:33238864
        • 78. Maheswari P, Harish S, Navaneethan M, Muthamizhchelvan C, Ponnusamy S, Hayakawa Y. Bio-modified TiO2 nanoparticles with Withania somnifera, Eclipta prostrata and Glycyrrhiza glabra for anticancer and antibacterial applications. Mater Sci Eng C Mater Biol Appl. 2020;108:110457. pmid:31924033
        • 79. Kukula-Koch W, Grabarska A, Łuszczki J, Czernicka L, Nowosadzka E, Gumbarewicz E, et al. Superior anticancer activity is demonstrated by total extract of Curcuma longa L. as opposed to individual curcuminoids separated by centrifugal partition chromatography. Phyther Res. 2018 May;32(5):933–42.
        • 80. Malaikozhundan B, Vaseeharan B, Vijayakumar S, Pandiselvi K, Kalanjiam MAR, Murugan K, et al. Biological therapeutics of Pongamia pinnata coated zinc oxide nanoparticles against clinically important pathogenic bacteria, fungi and MCF-7 breast cancer cells. Microb Pathog. 2017;104:268–77. pmid:28115262
        • 81. Chen G, Zhou D, Li X-Z, Jiang Z, Tan C, Wei X-Y, et al. A natural chalcone induces apoptosis in lung cancer cells: 3D-QSAR, docking and an in vivo/vitro assay. Sci Rep. 2017;7(1):10729. pmid:28878321
        • 82. Tsai Y. Social security income and the utilization of home care: Evidence from the social security notch. J Health Econ. 2015;43:45–55. pmid:26184382
        • 83. Varamini P, Soltani M, Ghaderi A. Cell cycle analysis and cytotoxic potential of Ruta graveolens against human tumor cell lines. Neoplasma. 2009;56(6):490–3. pmid:19728756
        • 84. Xie Y, Tang S, Qin Z, Yang C. Clinical Progress of Fruquintinib in Colorectal Cancer: An Overview. Pharmaceuticals (Basel). 2025;18(2):280. pmid:40006092
        • 85. Sargsyan K, Grauffel C, Lim C. How molecular size impacts RMSD applications in molecular dynamics simulations. J Chem Theory Comput. 2017;13(4):1518–24.
        • 86. Saha O, Siddiquee NH, Akter R, Sarker N, Bristi UP, Sultana KF, et al. Antiviral activity, pharmacoinformatics, molecular docking, and dynamics studies of Azadirachta indica against Nipah virus by targeting envelope glycoprotein: emerging strategies for developing antiviral treatment. Bioinform Biol Insights. 2024;18.
        • 87. Talukder MEK, Aktaruzzaman M, Siddiquee NH, Islam S, Wani TA, Alkahtani HM, et al. Cheminformatics-based identification of phosphorylated RET tyrosine kinase inhibitors for human cancer. Front Chem. 2024;12:1407331. pmid:39086985
        • 88. Saha S, Biswas P, Tareq MI, Rahman SM, Akter RS, Zilani MNH. Pharmacoinformatics, molecular dynamics simulation, and quantum mechanics calculation based phytochemical screening of Croton bonplandianum against breast cancer by targeting estrogen receptor-α (ERα). Appl Sci. 2024;14(21).
        • 89. Azmal M, Paul JK, Prima FS, Talukder OF, Ghosh A. An in silico molecular docking and simulation study to identify potential anticancer phytochemicals targeting the RAS signaling pathway. PLoS One. 2024;19(9):e0310637. pmid:39298437
        • 90. Bharadwaj S, Dubey A, Yadava U, Mishra SK, Kang SG, Dwivedi VD. Exploration of natural compounds with anti-SARS-CoV-2 activity via inhibition of SARS-CoV-2 Mpro. Brief Bioinform. 2021;22(2):1361–77. pmid:33406222
        • 91. Siddiquee NH, Hossain MI, Priya FM, Azam SB, Talukder MEK, Barua D, et al. Nature’s defense against emerging neurodegenerative threats: Dynamic simulation, PCA, DCCM identified potential plant-based antiviral lead targeting borna disease virus nucleoprotein. PLoS One. 2024;19(12):e0310802. pmid:39774339
        • 92. Holowatyj AN, Maude AS, Musa HS, Adamu A, Ibrahim S, Abdullahi A, et al. Patterns of Early-Onset Colorectal Cancer Among Nigerians and African Americans. JCO Glob Oncol. 2020;6:1647–55. pmid:33141623
        • 93. Ogbodo UC, Enejoh OA, Okonkwo CH, Gnanasekar P, Gachanja PW, Osata S. Computational identification of potential inhibitors targeting cdk1 in colorectal cancer. Front Chem. 2023;11:1–17.
        • 94. Cragg GM, Newman DJ. Natural products: a continuing source of novel drug leads. Biochim Biophys Acta. 2013;1830(6):3670–95. pmid:23428572
        • 95. Atanasov AG, Waltenberger B, Pferschy-Wenzig E-M, Linder T, Wawrosch C, Uhrin P, et al. Discovery and resupply of pharmacologically active plant-derived natural products: A review. Biotechnol Adv. 2015;33(8):1582–614. pmid:26281720
        • 96. Wang M, Liu X, Chen T, Cheng X, Xiao H, Meng X, et al. Inhibition and potential treatment of colorectal cancer by natural compounds via various signaling pathways. Front Oncol. 2022;12:956793. pmid:36158694
        • 97. Das A, Adhikari S, Deka D, Baildya N, Sahare P, Banerjee A. An updated review on the role of nanoformulated phytochemicals in colorectal cancer. Med. 2023;59(4).
        • 98. Gaieb Z, Parks CD, Chiu M, Yang H, Shao C, Walters WP, et al. D3R Grand Challenge 3: blind prediction of protein-ligand poses and affinity rankings. J Comput Aided Mol Des. 2019;33(1):1–18. pmid:30632055
        • 99. Meanwell NA. Improving drug candidates by design: a focus on physicochemical properties as a means of improving compound disposition and safety. Chem Res Toxicol. 2011;24(9):1420–56. pmid:21790149
        • 100. Tran TTV, Tayara H, Chong KT. Recent Studies of Artificial Intelligence on In Silico Drug Absorption. J Chem Inf Model. 2023;63(20):6198–211. pmid:37819031
        • 101. Chang Y, Hawkins BA, Du JJ, Groundwater PW, Hibbs DE, Lai F. A Guide to In Silico Drug Design. Pharmaceutics. 2022;15(1):49. pmid:36678678
        • 102. Al-Mohaya M, Mesut B, Kurt A, Çelik YS. In silico approaches which are used in pharmacy. J Appl Pharm Sci. 2024;14(4):225–39.