Integrated Computational Pipeline for the Identification of Novel PARP-1 Inhibitors: Hybrid Virtual Screening and Molecular Dynamics Simulations

Integrated Computational Pipeline for the Identification of Novel PARP-1 Inhibitors: Hybrid Virtual Screening and Molecular Dynamics Simulations | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Integrated Computational Pipeline for the Identification of Novel PARP-1 Inhibitors: Hybrid Virtual Screening and Molecular Dynamics Simulations Guan Wang, Jingjing Guo, Feng Xu, Mingjuan Ji This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4746837/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Despite the promising anticancer properties of PARP-1 inhibitors, their clinical use is hindered by side effects. It is crucial to explore new structural variants of these inhibitors to increase efficacy and minimize side effects, enhancing their clinical viability and therapeutic scope. In this study, we developed a virtual screening workflow that synergistically integrates Transfoxmol, KarmaDock, and PLANET with AutoDock Vina's capabilities. Through structural clustering, we identified ten potential PARP-1 inhibitors. Additionally, through molecular dynamics simulations and MM/PBSA, we elucidated the binding modes of compounds 1, 3, 6, and 9 with PARP-1, providing insights for drug development. Biological sciences/Cancer Biological sciences/Drug discovery Physical sciences/Chemistry/Theoretical chemistry/Computational chemistry Physical sciences/Chemistry/Theoretical chemistry/Molecular dynamics PARP-1 Hybrid Virtual Screening Molecular Dynamics Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 1. Introduction Poly(ADP-ribose) polymerase 1 (PARP-1) is a versatile enzyme pivotal in post-translational modifications, playing a key role in several biological functions, such as DNA repair[ 1 ]. PARP-1 is a prominent member of the PARP family, which encompasses 17 distinct members[ 2 ]. As depicted in Fig. 1 A, the structure of PARP-1 is comprising multiple domains: the N -terminal zinc finger domains (Zn1, Zn2, and Zn3), the BRCT domain, the WGR domain, and the C -terminal catalytic (CAT) domain[ 3 ]. The zinc finger domains Zn1 and Zn2, located at the N -terminus, are specialized in recognizing specific DNA structures. The BRCT domain primarily facilitates automodification[ 4 ]. The CAT domain, which includes the HD and ART subdomains, is crucial for the enzyme's activity, mediating the addition of ADP-ribose polymers to target proteins, thereby enhancing the DNA repair mechanism as illustrated in Fig. 1 B[ 5 ]. During cell growth, DNA damage inevitably occurs due to internal and external factors[ 7 ]. Such damage can lead to cell cycle arrest or genomic instability, both of which are key characteristics that contribute to cellular transformation into cancer. PARP-1 inhibitors, a class of small molecule drugs targeting PARP-1, act through two primary mechanisms: synthetic lethality and immune activation. Synthetic lethality exploits DNA repair deficiencies in tumor cells. For instance, certain cancer cells, such as those with BRCA mutations found in breast and ovarian cancers, lack the homologous recombination repair (HRR) pathway, which is necessary for repairing double-strand breaks (DSBs). While normal cells repair DSBs through HRR, cancer cells are unable to do so[ 8 , 9 ]. As a result, PARP-1 inhibitors block the repair of single-strand breaks (SSBs), causing these breaks to evolve into DSBs during DNA replication, leading to the death of tumor cells while sparing normal cells. On the other hand, immune activation by PARP-1 inhibitors induces DNA damage that results in tumor cells producing more mutations and neoantigens, allowing them to be recognized and attacked by the immune system, thus achieving anticancer effects[ 10 ]. The first-generation PARP-1 inhibitors, based on the nicotinamide structure, incorporated electron-donating groups or bioisosteres to develop analogs like 3-aminobenzamide (3-AB), enhancing intermolecular interactions and inhibition efficacy. Second-generation PARP-1 inhibitors have improved upon drug design and structural changes to enhance potency and therapeutic outcomes. As shown in Fig. 2 , since the market launch of the first PARP-1 inhibitor, Olaparib, in 2014, there have been significant breakthroughs in the development of PARP-1 inhibitors. Subsequently, the FDA has approved several others, including Niraparib, Rucaparib, and Talazoparib, highlighting the critical role of PARP-1 inhibitors in cancer treatment. Currently, numerous PARP-1 inhibitors are undergoing clinical trials. For instance, Simmiparib, developed by the Shanghai Institute of Pharmaceutical Research, substitutes the triazine ring in the structure of Olaparib with a triazole ring, achieving an inhibition rate five times higher than Olaparib with an IC 50 of 0.74 nM[ 11 ]. Sun et al. have synthesized derivatives containing phthalazin-1(2H)-one, where YCH1899 showed significant antiproliferative activity against Olaparib and Talazoparib-resistant cells with IC 50 values of 0.89 and 1.13 nM, respectively[ 12 ]. Despite the promising anticancer activity of PARP-1 inhibitors, they still exhibit side effects that limit their clinical applications. Therefore, it remains essential to explore new structural types of PARP-1 inhibitors to enhance efficacy and reduce side effects, thereby improving their clinical applicability and expanding their therapeutic potential. In conventional computationally assisted drug design (CADD), the screening framework depends on the structures of proteins or ligands, with two main types: ligand-based (LBVS) and structure-based (SBVS) screening. LBVS concentrates on the structure of ligands but may neglect critical details of the target proteins, which can result in suboptimal binding effectiveness. On the other hand, SBVS uses protein structures but is often hampered by their limited availability and the difficulties in managing large-scale databases, which can lead to increased time and computational demands. Additionally, CADD faces challenges in precisely predicting how compounds will interact with target proteins. To overcome these limitations, researchers are increasingly incorporating artificial intelligence (AI) into CADD processes to boost both the accuracy and efficiency of predictions. AI technologies, especially deep learning and attention mechanisms, are adept at capturing and processing complex molecular features and interactions. This enhances the identification of active molecules and optimizes the screening workflow. For example, the Deep Docking platform employs iterative methods to improve the precision of predictions and increase the overall efficiency of the screening process, showcasing how AI can significantly expedite the drug discovery process by refining and accelerating CADD operations. In this study, we used hybrid virtual screening workflow to identified novel scaffold PAPR-1inhibitor molecules. The process involves multiple stages, including for Transfoxmol, KarmaDock, PLANET, and Vina for generated Conformation. We also applied MM/PBSA to evaluate further binding potential between candidate compounds and PARP-1. After several iterations of the workflow, our approach achieved promising results, with 10 compounds standing out. Furthermore, we selected three candidate molecules and performed a detailed study of the binding interactions between the compounds and their targets using molecular dynamics simulations. Therefore, we plan to conduct further molecular structure modification and optimization in the laboratory. 2. Computational details and methods 2.1. Choice of PARP-1 structure and preparation In our research, we conducted a thorough exploration of the UniProt database[ 13 ], specifically targeting the PARP-1_Human catalytic domain (662–1011) with the UniProt ID P09874. From this inquiry, we obtained 55 X-ray co-crystal structures from the RCSB Protein Data Bank[ 14 ]. These structures were subjected to detailed analysis using the Structural Analysis and Verification Server (SAVES) v6.0, incorporating both the PROCHECK[ 15 , 16 ] and ERRAT[ 17 ] modules for validation. SAVES, which can be accessed at https://saves.mbi.ucla.edu/ (as of January 13, 2024), provides a comprehensive suite for assessing the structural integrity and reliability of protein models. For the virtual screening, we specifically prepared the 7KK5 structure. Utilizing PyMOL[ 18 ], we meticulously removed water molecules and added hydrogen atoms to the structure. 2.2. Database preparation In our study, we developed a virtual screening workflow centered on the Topscience database ( https://www.tsbiochem.com/ ), which comprises approximately 13 million molecules. The integrity and quality of our database are critical for the success of the screening process. To achieve this, we employed the RDKit[ 19 ] to implement a rigorous data preprocessing protocol. Our approach began with the identification and removal of duplicate molecular structures, ensuring that each molecule in our database is unique. This was followed by the parsing of each molecule, represented as a SMILES string, into a molecular structure object using RDKit. We discarded entries with invalid SMILES strings, indicating parsing failures. Subsequent steps involved the removal of salt components from molecules with RDKit's SaltRemover and the neutralization of molecular charges using the Uncharger function. These steps are crucial for isolating the principal active component of each molecule, making the dataset more uniform and facilitating consistent comparisons and analyses. Additionally, we conducted a boron valence verification to exclude molecules with boron atoms exhibiting explicit valences greater than three, as these do not conform to standard chemical validity. Finally, we standardized the molecular structures by converting them into the universally recognized SMILES format, preparing them for subsequent computational analyses. 2.3. Virtual screening Transfoxmol TransFoxMol (available on https://github.com/gaojianl/TransFoxMol )[ 20 ] was conceived to augment the artificial intelligence's understanding of the intricate relationships between molecular structures and their properties. This method is notable for integrating a multi-scale 2D molecular environment into a cohesive framework that combines a graph neural network with a Transformer architecture. It uniquely utilizes pre-existing chemical maps to fine-tune its attention mechanism, thereby achieving a level of precision and efficiency beyond what current methodologies offer. The foundation for training this model was a dataset sourced from the ChEMBL database[ 21 , 22 ] as of January 12, 2024. This dataset was meticulously processed to exclude duplicate entries and those lacking essential information such as labels or SMILES notation. It was further enhanced by converting original IC 50 values to pIC 50 values, thus providing a more accurate reflection of molecular activity, represented as -lnIC 50 = pIC 50 . The experimental process commenced with the preparation of the PARP-1 dataset for regression analysis using the molnetdata.py script, followed by hyperparameter optimization executed through the ‘run.py search’. The model's training was precisely outlined with parameters set for regression analysis on the PARP-1 dataset (--task reg), utilizing a CUDA device, and specified settings including a batch size of 32, a total of 50 training epochs, a learning rate of 0.0005, both validation and test rates at 0.1, a seed value of 426, conducted over three folds. Additionally, the model featured a dropout rate of 8.85, eight attention heads, two attention layers, an output dimensionality of 128, and incorporated four distance thresholds (--D 4). The primary metric for assessing the model's performance was the Root Mean Square Error (RMSE), with the best validation RMSE recorded at 0.7569 and the best test RMSE at 0.8109. The testing phase reported a loss of 0.0974, a mean absolute error (MAE) of 0.4337, an RMSE of 0.5971, and an R 2 score of 0.7861. For the prediction stage, the model, trained on the PARP-1 dataset and encapsulated as PARP-1reg_44_0.7569000124931335.pkl, was applied to the prepared Topscience database to facilitate further molecular activity predictions. KarmaDock We utilized KarmaDock, a cutting-edge deep learning framework designed for ligand docking, which is publicly available on GitHub ( https://github.com/schrojunzhang/KarmaDock )[ 23 ]. This method encapsulates functionalities for expedited docking, generation and refinement of binding poses, and evaluation of binding affinities. KarmaDock's architecture is structured into a three-stage model: (1) It employs encoders to capture the protein and ligand's intramolecular interaction features. (2) It utilizes E(n) equivariant graph neural networks complemented with self-attention mechanisms to refine the ligand's pose, taking into account both protein-ligand and intraligand interactions. This stage is augmented by a post-processing step to guarantee the chemical validity of the generated structures. (3) It incorporates a mixture density network to quantitatively assess the binding affinity. In our experimental setup, the ligand was represented in '.smi' format for input purposes. The target protein, identified by PDB ID: 7kk5, along with its crystal-bound ligand, was retrieved from the RCSB Protein Data Bank[ 14 ]. Docking simulations were executed utilizing the virtual_screening_pipeline.py script. PLANET The PLANET model[ 24 ], accessible via https://github.com/ComputArtCMCG/PLANET/ , was utilized, incorporating the 3D structure of a target protein's binding pocket represented as a graph and the 2D chemical structure of the ligand in ‘.sdf’ format as its primary inputs. The model underwent an extensive training regimen following a tri-objective approach, aimed at assessing protein-ligand binding affinity, mapping out protein-ligand interaction sites, and creating the ligand distance matrix. For our experiments, the ligand's chemical structure, provided in ‘.sdf’ format, served as the input. The protein of interest, bearing the PDB ID: 7kk5, and its crystal-bound ligand were sourced from the RCSB database[ 14 ]. The docking analysis was performed using the PLANET_run.py script. AutoDock Vina AutoDock Vina (version 1.2.3)[ 25 , 26 ] was deployed for ligand preparation, utilizing prepare_ligand4.py, and for generating the configuration file through prepare_pdb_split_alt_confs.py, which led to the creation of configure.txt. This file specified parameters such as: center_x = -36.91, center_y = 5.89, center_z = -7.45; size_x = 30, size_y = 30, size_z = 30; energy_range = 3; exhaustiveness = 48; num_modes = 20. Following this, prepare_receptor4.py was applied for receptor preparation, which included the removal of water molecules and the confirmation of hydrogen atoms' presence. The docking procedure was subsequently executed using Vina. Filtered and Cluster In the final stage of virtual screening, our methodology incorporated a suite of Python libraries for data manipulation and analysis, including RDKit for cheminformatics, Matplotlib[ 27 ] and Pandas[ 28 ] for data visualization and manipulation, NumPy[ 29 ] for numerical operations, SciPy[ 30 ] for distance calculations, and sklearn for clustering algorithms. Our process commenced with loading a dataset of molecular docking scores, derived from both Vina and PLANET analyses, from a CSV file. This dataset was then refined to isolate molecules exhibiting PLANET scores above 10 and Vina scores below − 10, resulting in a cohort of 191 molecules targeted for clustering. The clustering phase involved generating Morgan fingerprints[ 31 ] for each molecule. These fingerprints are structured as fixed-length bit vectors, encoding the presence or absence of particular molecular substructures, thereby serving as a compact representation of molecular features. Subsequent steps entailed calculating pairwise Jaccard distances among these fingerprints to assess chemical dissimilarities between molecules. The Jaccard distance[ 32 ], a metric of dissimilarity between two sets, facilitated the quantitative analysis of chemical diversity within our dataset. We then applied agglomerative hierarchical clustering to this distance matrix. This clustering technique progressively merges pairs of clusters based on their proximity, utilizing precomputed Jaccard distances as the affinity metric and adopting an average linkage strategy to gauge distances between clusters. Through this method, each molecule was assigned to one of ten predetermined clusters, reflecting their chemical similarities. 2.4. Molecular dynamics simulation In our research, we utilized molecular dynamics simulations to investigate protein-ligand interactions, employing GROMACS[ 33 ] (version 2020.7) and structures from AutoDock Vina[ 34 ]. We formatted protein and ligand structures in PDB, refining proteins with pdbfixer ( https://github.com/openmm/pdbfixer ) and using the amber99sb-ildn force field. Our simulations took place in a 1.5 nm cubic box using the tip3p water model. For ligand-containing systems, we derived parameters for both ligands and receptors, converting ligand structures to mol2 format with Open Babel, and parameterizing with ACPYPE (version 1.3.0)[ 35 ]. We created GROMACS-compatible receptor topology files with pdb2gmx and integrated them with ligand topologies into a unified file. We applied the -DFLEXIBLE option before starting energy minimization with the Conjugate Gradient algorithm, interspersed with steepest descent corrections to prevent local minima, ceasing minimization when forces dropped below 1000 kJ/mol/nm. For the NVT equilibration phase, we used a leap-frog integrator over 500 ps with data recording at 1.0 ps intervals. Spatial searches utilized the grid method, with long-range electrostatics managed via the Particle Mesh Ewald method. Temperature was controlled using the Berendsen thermostat at 300 K. The NPT phase followed with the same settings, and pressure controlled by the Berendsen method at 1.0 bar isotropic coupling. Our production phase spanned 150 ns, capturing data every 100 ps, with all hydrogen bonds constrained by the LINCS algorithm and thermodynamic parameters managed using the V-rescale and Parrinello-Rahman methods. 2.5. MM/PBSA We conducted MM-PBSA calculations by employing a custom bash script designed specifically for processing GROMACS[ 33 ] (version 2020.7) molecular dynamics simulations and Adaptive Poisson-Boltzmann Solver (APBS) (version 3.0.0)[ 36 ] for calculating electrostatic potentials. This script facilitated the efficient preparation of input files by automating the extraction of trajectory data (.xtc), topology (.tpr), and index files (*.ndx). The resources were made available through Jerkwin's gmxtools repository on GitHub ( https://github.com/Jerkwin/gmxtools ) [ 37 ]. The streamlined procedure included preparing input files from GROMACS outputs, setting up APBS[ 36 ] calculations, and estimating binding free energies. For the segregation of ligands and proteins, specific index groups were identified. In the configuration file (configure.txt), we meticulously defined parameters, including those for electrostatic and van der Waals interactions, the Debye-Hückel screening method for MM (Coulombic) energies, and criteria for selecting residues for the calculations[ 38 ]. These parameters were crucial for ensuring the accuracy of simulations, encompassing solvent models and boundary conditions, thus facilitating precise electrostatic potential evaluation by APBS. Energy calculations, encompassing molecular mechanics, polar, and non-polar solvation energies, were performed, leading to the estimation of binding free energies. The analysis was based on trajectory data (.xtc), topology information (.tpr), and index files (*.ndx) derived from the final 50 ns of 150 ns molecular dynamics simulation. 3. Results and discussion 3.1. The evaluation of PARP-1 structures In our quest to identify the optimal PARP-1 protein structure for our study, we initiated our search within the UniProt database[ 13 ] for the PARP-1 enzyme specific to Homo sapiens (Human), bearing the UniProt ID P09874 ( https://www.uniprot.org , access date: 2024-1-13), and identified 55 PARP1_HUMAN structures containing catalytic domain (662–1011). Our selection criteria excluded any structures exhibiting mutations or those with a resolution exceeding 2Å, to ensure high fidelity in our computational analyses (Fig. 3 A, Table 1 ). Further scrutiny was directed towards four remaining structures, evaluated using the SAVES v6.0 platform, a comprehensive tool for assessing structural integrity and reliability. This analysis encompassed several aspects; firstly, VERIFY 3D inspection was applied to assess the compatibility of each structure's amino acid sequence with its 3D structure, with all nine structures surpassing the threshold of 80 points (Fig. 3 B, Table 1 ). Subsequent examination employed the PROCHECK module, which gauges the stereochemical quality of the protein structures. Among the structures analyzed, 7KK4 was noted for having less than 90% of its residues in the most preferred regions, leading to its exclusion from further consideration (Fig. 3 C, Table 1 ). Additionally, the ERRAT module, designed for the evaluation of crystallography-derived protein structures, provided an overall quality factor for each structure (Fig. 3 D, Table 1 ). Notably, structure 7KK2 emerged as the optimal choice based on this metric; however, it was ultimately deemed unsuitable due to the absence of small molecule compounds within its structure. Considering both resolution and PROCHECK scores, we concluded that structure 7KK5 represented the best candidate for the foundation of our virtual screening efforts in this study. Table 1 PARP1 structure Quality Metrics and Ramachandran Plot Statistics PDB ID Mutation Resolution Released Overall quality factor VERIFY 3D Ramachandran plot 4ZZZ No 1.90 Å 2015-08-12 95.0966 Pass 91.9% core 8.1% allow 0.0% gener 0.0% disal 7KK2 No 1.70 Å 2021-01-06 97.0458 Pass 92.4% core 7.6% allow 0.0% gener 0.0% disa 7KK4 No 1.96 Å 2021-01-06 95.3632 Pass 89.8% core 10.0% allow 0.2% gener 0.0% disal 7KK5 No 1.70 Å 2021-01-06 96.5839 Pass 92.2% core 7.8% allow 0.0% gener 0.0% disal 3.2. Hybrid Virtual screening In our pursuit to identify potent PARP-1 inhibitors, we adopted an integrative approach, combining AI-driven screening methodologies with conventional computer-aided drug design (CADD) tools (Fig. 4 A). Initially, we leveraged TransFoxMol, a cutting-edge tool recognized for its innovative integration of multi-scale 2D molecular environments within a graph neural network and Transformer architecture framework. TransFoxMol excels in correlating molecular structures with their properties, thereby facilitating an efficient initial screening of the PAPR1 activity dataset sourced from the ChEMBL database. Through this initial phase, we utilized TransFoxMol to screen the extensive Topscience database (total of 13 million molecules), filtering for molecules predicted with a score above 7.5, resulting in 189,480 candidates. For the subsequent round of screening, we employed KarmaDock, an AI-based docking tool renowned for its precision in rapidly predicting the binding strength of protein-ligands. By setting a threshold Karma score of above 80, we further refined our selection to 3,471 molecules. The third phase of our screening process involved the use of PLANET, another AI-based tool designed to predict protein-ligand binding affinity with high accuracy. Despite the strengths of AI in predicting molecular activity and affinity, these tools often fall short in accurately generating 3D conformations of compounds. To address this limitation, we incorporated AutoDock Vina, a traditional molecular docking tool, to enhance the precision of compound conformations. Focusing on molecules that achieved a PLANET score above 10 and a Vina score below − 10 kcal/mol, we distilled our list to 192 compounds. To assess the structural diversity within our dataset, we utilized Morgan molecular fingerprints, a method allowing for the detailed categorization of molecules into one of ten pre-established clusters reflecting their chemical similarity. This clustering enabled us to distill our findings further to ten unique molecules (Fig. 4 B and Table 2 ). The selection criterion for further analysis was primarily based on achieving the highest Vina docking scores, from which compounds 1, 3, 6, and 9 were chosen for advanced Molecular Dynamics (MD) simulations. This approach, a synergy of hybrid virtual screening and structural clustering, integrates the predictive prowess of artificial intelligence with the robust methodologies of CADD technologies. By doing so, we not only streamline the identification process of promising PARP-1 inhibitor candidates but also furnish future researchers with novel PARP-1 scaffolds. Table 2 Each stage score of the final selected 10 small molecules. Entry InChIKey Transfoxmol KarmaDock PLANET Vina MM/PBSA (kcal/mol) Compd. 1 InChIKey = ORBZKIMBPNECHS-UHFFFAOYSA-N 7.715 93.262 16.094 12.887 -24.013 Compd. 2 InChIKey = YMJJZAYPAWMBMO-PTTDRDKLNA-N 7.842 82.217 11.283 10.796 - Compd. 3 InChIKey = ULUVBBCDTWTZLD-UHFFFAOYSA-N 8.541 94.830 12.847 14.211 -40.081 Compd. 4 InChIKey = HYUHVOLQYDHZHQ-UHFFFAOYSA-N 7.587 85.068 10.942 11.482 - Compd. 5 InChIKey = DJIKKGWKXJGKRI-WYOOIXGGSA-N 7.503 108.382 15.740 10.798 - Compd. 6 InChIKey = DSYYGKZYBMLNEM-FLFKKZLDSA-N 7.778 88.667 11.766 11.573 -20.806 Compd.7 InChIKey = NDIZSXKXYRQMAU-SXGWCWSVSA-N 7.640 81.205 12.289 10.836 - Compd. 8 InChIKey = ANAHWPGZGLJKPD-UHFFFAOYNA-N 7.589 80.588 10.901 10.796 - Compd. 9 InChIKey = RVNGBBQZELJWBB-UHFFFAOYSA-N 8.038 93.073 12.382 11.851 -12.880 Compd.10 InChIKey = GCVOEPYFYISOOG-UHFFFAOYSA-N 7.597 85.930 11.910 11.491 - 3.3. Determination of structural stability and dynamical of the complexes The determination of structural stability and dynamics of complexes is a critical phase in the computational analysis of protein-ligand interactions. This step involves conducting Molecular Dynamics (MD) simulations, a computational method that enables the exploration of the physical movements of atoms and molecules over time. Here, our objective is to evaluate the behavior of selected PARP-1 inhibitor complexes under simulated physiological conditions, providing insights into their conformational stability, flexibility, and overall dynamic behavior. As depicted in Figs. 5 A and B, the trajectories of backbones and ligands reach equilibrium in the final 10 ns. Figure 5 C illustrates the RMSF plot for PARP-1, demonstrating amino acid residue fluctuations ranging from 0.1 Å to 0.6 Å. Furthermore, an analysis of the Gibbs Free Energy landscape (Figs. 5 D, G, J, M) identified the most energetically favorable compounds 1, 3, 6, 9, and PARP-1. Notably, compound 6 forms multiple hydrogen bond interactions with Arg878, Gly863, and Ser904 of PARP-1, while compound 6 interacts via a hydrogen bond with Arg865 of PARP-1. Additionally, compound 9 forms multiple hydrogen bond interactions with Arg878, Asp770, His862, Ser864, Gly863, Ser904, Asp766, and Met980 of PARP-1. 3.4. Binding free energy calculations In the realm of drug discovery, accurately assessing the binding free energy is essential for evaluating the efficacy and specificity of drug candidates. The Molecular Mechanics/Poisson-Boltzmann Surface Area (MM/PBSA) method stands out as a crucial technique in this process due to its optimal balance between computational intensity and accuracy. In our study, MM/PBSA analyses were performed using data from molecular dynamics (MD) simulation trajectories. These analyses shed light on the contributions of individual amino acid residues to the overall binding free energy in various PARP-1/compound complexes. As depicted in Fig. 5 , negative values indicate beneficial interactions that enhance the stability of the complex. For instance, the binding free energy of compound 1 with PARP-1 was calculated to be -24.013 kcal/mol, with significant contributions from residues like PRO881, TYR907, and TYR889 (Fig. 6 A, Table 2 ). Similarly, compound 3 demonstrated a binding free energy of -40.081 kcal/mol with notable contributions from TYR896, TYR907, and GLU763 (Fig. 6 B, Table 2 ). Compound 6 showed a binding free energy of -20.806 kcal/mol, influenced significantly by TYR896, TYR907, and GLU889 (Fig. 6 C, Table 2 ). Lastly, a repeat analysis for compound 6/PARP-1 presented a binding free energy of -12.880 kcal/mol with contributions from TYR907, TYR896, and ASN868 being particularly impactful (Fig. 6 D, Table 2 ). 4. Conclusions Despite the promising anticancer activity of PARP-1 inhibitors, they still exhibit side effects that limit their clinical applications. Therefore, exploring new structural types of PARP-1 inhibitors to enhance efficacy and reduce side effects remains essential, thereby improving their clinical applicability and expanding their therapeutic potential. In this study, we conducted a comprehensive screening process aimed at identifying potential PARP-1 inhibitors. Our methodology integrated Transfoxmol, KarmaDock, and PLANET with AutoDock Vina, utilizing it for rescoring and generating conformations. This combined approach successfully identified ten potent PARP-1 inhibitors from the Topscience database. In the final round of screening, we employed clustering to obtain ten structurally diverse compounds, providing insights for future PARP-1 inhibitor design. Furthermore, molecular dynamics simulations and MM/PBSA were performed to precisely elucidate the binding kinetics of compounds 1, 3, 6, and 9 with PARP-1, offering a clear understanding of their interaction mechanisms. An analysis of the Gibbs Free Energy landscape identified the most energetically favorable compounds 1, 3, 6, 9, and PARP-1. Notably, compound 6 formed multiple hydrogen bond interactions with Arg878, Gly863, and Ser904 of PARP-1, while compound 6 interacted via a hydrogen bond with Arg865 of PARP-1. Additionally, compound 9 formed multiple hydrogen bond interactions with Arg878, Asp770, His862, Ser864, Gly863, Ser904, Asp766, and Met980 of PARP-1. Declarations Acknowledgments This work was supported by the [Taizhou Science and technology plan] under Grant [number 22ywa61]; [Zhejiang soft Science research program] under Grant [number 2024C35015]; and [Taizhou Vocational and Technical College doctoral research fund] under Grant [number 14071191]. Disclosure statement There are no conflicts of interest to declare. Data availability If you need research data related to the article, please request it from the corresponding author. References Ray Chaudhuri, A.; Nussenzweig, A. The Multifaceted Roles of PARP1 in DNA Repair and Chromatin Remodelling. Nat Rev Mol Cell Biol 2017, 18 , 610–621, doi: 10.1038/nrm.2017.53 . Alemasova, E.E.; Lavrik, O.I. Poly(ADP-Ribosyl)Ation by PARP1: Reaction Mechanism and Regulatory Proteins. Nucleic Acids Research 2019, 47 , 3811–3827, doi: 10.1093/nar/gkz120 . Langelier, M.-F.; Pascal, J.M. PARP-1 Mechanism for Coupling DNA Damage Detection to Poly(ADP-Ribose) Synthesis. Current Opinion in Structural Biology 2013, 23 , 134–143, doi: 10.1016/j.sbi.2013.01.003 . Fu, X.; Li, P.; Zhou, Q.; He, R.; Wang, G.; Zhu, S.; Bagheri, A.; Kupfer, G.; Pei, H.; Li, J. Mechanism of PARP Inhibitor Resistance and Potential Overcoming Strategies. Genes & Diseases 2024, 11 , 306–320, doi: 10.1016/j.gendis.2023.02.014 . Kanev, P.-B.; Atemin, A.; Stoynov, S.; Aleksandrov, R. PARP1 Roles in DNA Repair and DNA Replication: The Basi(c)s of PARP Inhibitor Efficacy and Resistance. Seminars in Oncology 2024, 51 , 2–18, doi: 10.1053/j.seminoncol.2023.08.001 . Ryan, K.; Bolanos, B.; Smith, M.; Palde, P.B.; Cuenca, P.D. Structure of the Catalytic Domain of PARP1 in Complex with Niraparib. 2021, doi: https://doi.org/10.2210/pdb7KK5/pdb . DHANYA K. NAMBIAR, 3, DEEPALI MISHRA, RANA P. SINGH Targeting DNA Repair for Cancer Treatment: Lessons from PARP Inhibitor Trials. Oncology Research 2023, 31 , 405–421, doi: 10.32604/or.2023.028310 . Zhao, Y.; Zhang, L.-X.; Jiang, T.; Long, J.; Ma, Z.-Y.; Lu, A.-P.; Cheng, Y.; Cao, D.-S. The Ups and Downs of Poly(ADP-Ribose) Polymerase-1 Inhibitors in Cancer Therapy–Current Progress and Future Direction. European Journal of Medicinal Chemistry 2020, 203 , 112570, doi: 10.1016/j.ejmech.2020.112570 . Lord, C.J.; Tutt, A.N.J.; Ashworth, A. Synthetic Lethality and Cancer Therapy: Lessons Learned from the Development of PARP Inhibitors. Annual Review of Medicine 2015, 66 , 455–470. Abbotts, R.; Wilson, D.M. Coordination of DNA Single Strand Break Repair. Free Radical Biology and Medicine 2017, 107 , 228–244, doi: 10.1016/j.freeradbiomed.2016.11.039 . Ye, N.; Chen, C.-H.; Chen, T.; Song, Z.; He, J.-X.; Huan, X.-J.; Song, S.-S.; Liu, Q.; Chen, Y.; Ding, J.; et al. Design, Synthesis, and Biological Evaluation of a Series of Benzo[de][1,7]Naphthyridin-7(8H)-Ones Bearing a Functionalized Longer Chain Appendage as Novel PARP1 Inhibitors. J. Med. Chem. 2013, 56 , 2885–2903, doi: 10.1021/jm301825t . Sun, Y.; Yang, H.; Yuan, J.; Wang, L.; Song, S.; Chen, R.; Bao, X.; Jia, L.; Yang, T.; Zhang, X.; et al. YCH1899, a Highly Effective Phthalazin-1(2H)-One Derivative That Overcomes Resistance to Prior PARP Inhibitors. J. Med. Chem. 2023, 66 , 12284–12303, doi: 10.1021/acs.jmedchem.3c00821 . The UniProt Consortium UniProt: The Universal Protein Knowledgebase in 2023. Nucleic Acids Research 2023, 51 , D523–D531, doi: 10.1093/nar/gkac1052 . RCSB Protein Data Bank: Efficient Searching and Simultaneous Access to One Million Computed Structure Models Alongside the PDB Structures Enabled by Architectural Advances. Journal of Molecular Biology 2023, 435 , 167994, doi: 10.1016/j.jmb.2023.167994 . Laskowski, R.A.; Rullmannn, J.A.; MacArthur, M.W.; Kaptein, R.; Thornton, J.M. AQUA and PROCHECK-NMR: Programs for Checking the Quality of Protein Structures Solved by NMR. J Biomol NMR 1996, 8 , 477–486, doi: 10.1007/BF00228148 . Laskowski, R.A.; MacArthur, M.W.; Moss, D.S.; Thornton, J.M. PROCHECK: A Program to Check the Stereochemical Quality of Protein Structures. J Appl Cryst 1993, 26 , 283–291, doi: 10.1107/S0021889892009944 . Colovos, C.; Yeates, T.O. Verification of Protein Structures: Patterns of Nonbonded Atomic Interactions. Protein Sci 1993, 2 , 1511–1519, doi: 10.1002/pro.5560020916 . Schrödinger, LLC The PyMOL Molecular Graphics System, Version 1.8 2015. Landrum, G. RDKit: A Software Suite for Cheminformatics, Computational Chemistry, and Predictive Modeling 2019. Gao, J.; Shen, Z.; Xie, Y.; Lu, J.; Lu, Y.; Chen, S.; Bian, Q.; Guo, Y.; Shen, L.; Wu, J.; et al. TransFoxMol: Predicting Molecular Property with Focused Attention. Briefings in Bioinformatics 2023, 24 , bbad306, doi: 10.1093/bib/bbad306 . Davies, M.; Nowotka, M.; Papadatos, G.; Dedman, N.; Gaulton, A.; Atkinson, F.; Bellis, L.; Overington, J.P. ChEMBL Web Services: Streamlining Access to Drug Discovery Data and Utilities. Nucleic Acids Res 2015, 43 , W612–W620, doi: 10.1093/nar/gkv352 . Gaulton, A.; Hersey, A.; Nowotka, M.; Bento, A.P.; Chambers, J.; Mendez, D.; Mutowo, P.; Atkinson, F.; Bellis, L.J.; Cibrián-Uhalte, E.; et al. The ChEMBL Database in 2017. Nucleic Acids Research 2017, 45 , D945–D954, doi: 10.1093/nar/gkw1074 . Zhang, X.; Zhang, O.; Shen, C.; Qu, W.; Chen, S.; Cao, H.; Kang, Y.; Wang, Z.; Wang, E.; Zhang, J.; et al. Efficient and Accurate Large Library Ligand Docking with KarmaDock. Nat Comput Sci 2023, 3 , 789–804, doi: 10.1038/s43588-023-00511-5 . Zhang, X.; Gao, H.; Wang, H.; Chen, Z.; Zhang, Z.; Chen, X.; Li, Y.; Qi, Y.; Wang, R. PLANET: A Multi-Objective Graph Neural Network Model for Protein–Ligand Binding Affinity Prediction. J. Chem. Inf. Model. 2023, acs.jcim.3c00253 , doi:10.1021/acs.jcim.3c00253. Eberhardt, J.; Santos-Martins, D.; Tillack, A.F.; Forli, S. AutoDock Vina 1.2.0: New Docking Methods, Expanded Force Field, and Python Bindings. J. Chem. Inf. Model. 2021, 61 , 3891–3898, doi: 10.1021/acs.jcim.1c00203 . Trott, O.; Olson, A.J. AutoDock Vina: Improving the Speed and Accuracy of Docking with a New Scoring Function, Efficient Optimization, and Multithreading. Journal of Computational Chemistry 2010, 31 , 455–461, doi: 10.1002/jcc.21334 . Hunter, J.D. Matplotlib: A 2D Graphics Environment. Computing in Science & Engineering 2007, 9 , 90–95, doi: 10.1109/MCSE.2007.55 . team, T. pandas development Pandas-Dev/Pandas: Pandas 2024. Harris, C.R.; Millman, K.J.; Van Der Walt, S.J.; Gommers, R.; Virtanen, P.; Cournapeau, D.; Wieser, E.; Taylor, J.; Berg, S.; Smith, N.J.; et al. Array Programming with NumPy. Nature 2020, 585 , 357–362, doi: 10.1038/s41586-020-2649-2 . Virtanen, P.; Gommers, R.; Oliphant, T.E.; Haberland, M.; Reddy, T.; Cournapeau, D.; Burovski, E.; Peterson, P.; Weckesser, W.; Bright, J.; et al. SciPy 1.0: Fundamental Algorithms for Scientific Computing in Python. Nat Methods 2020, 17 , 261–272, doi: 10.1038/s41592-019-0686-2 . Morgan, H.L. The Generation of a Unique Machine Description for Chemical Structures-A Technique Developed at Chemical Abstracts Service. J. Chem. Doc. 1965, 5 , 107–113, doi: 10.1021/c160017a018 . Hancock, J.M. Jaccard Distance (Jaccard Index, Jaccard Similarity Coefficient). In Dictionary of Bioinformatics and Computational Biology ; 2014 ISBN 978-0-471-65012-6. Abraham, M.J.; Murtola, T.; Schulz, R.; Páll, S.; Smith, J.C.; Hess, B.; Lindahl, E. GROMACS: High Performance Molecular Simulations through Multi-Level Parallelism from Laptops to Supercomputers. SoftwareX 2015, 1 , 19–25, doi: 10.1016/j.softx.2015.06.001 . Trott, O.; Olson, A. Software News and Update AutoDock Vina: Improving the Speed and Accuracy of Docking with a New Scoring Function, Efficient Optimization, and Multithreading. JOURNAL OF COMPUTATIONAL CHEMISTRY 2010, 31 , 455–461, doi: 10.1002/jcc.21334 . Sousa da Silva, A.W.; Vranken, W.F. ACPYPE - AnteChamber PYthon Parser interfacE. BMC Research Notes 2012, 5 , 367, doi: 10.1186/1756-0500-5-367 . Jurrus, E.; Engel, D.; Star, K.; Monson, K.; Brandi, J.; Felberg, L.E.; Brookes, D.H.; Wilson, L.; Chen, J.; Liles, K.; et al. Improvements to the APBS Biomolecular Solvation Software Suite. Protein Science 2018, 27 , 112–128, doi: 10.1002/pro.3280 . Li, J. Gmxtools 2022. Genheden, S.; Ryde, U. The MM/PBSA and MM/GBSA Methods to Estimate Ligand-Binding Affinities. Expert Opinion on Drug Discovery 2015, 10 , 449–461, doi: 10.1517/17460441.2015.1032936 . Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-4746837","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":337863010,"identity":"181f4b92-f4d6-4f08-ac75-5b272229a161","order_by":0,"name":"Guan Wang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA6ElEQVRIiWNgGAWjYDADfjjrALFaJNtgqonWYnCMWC3yEcnPHvPU3LHbfL/9mfSHGgY5vhsJjJ8L8GgxvJFmbsxz7FnytmMMaRIHjjEYS95IYJaegU/LjAQz6Ry2w8lmxxgOGxxsYEjccCOBjZkHr5b0b9I5/w4nG7cxNoO01BPUIi+RYyad23bYzoCNmfEBUEuCASEtBjxvyqT/9h1OkDiWxvjgzDEJw5lnHjZL47WlPX2b5Ixvh+35m48/OFBRYyPPdzz54Ge8thyA0IkNEFoCiBkb8GgA2gKVtserahSMglEwCkY2AADDr06sKx2n8QAAAABJRU5ErkJggg==","orcid":"","institution":"Taizhou Vocational and Technical College, Chemical Pharmaceutical Research Institute","correspondingAuthor":true,"prefix":"","firstName":"Guan","middleName":"","lastName":"Wang","suffix":""},{"id":337863011,"identity":"e01b6db8-87a5-4b67-bf2c-3384e4c394ad","order_by":1,"name":"Jingjing Guo","email":"","orcid":"","institution":"Taizhou Vocational and Technical College, Chemical Pharmaceutical Research Institute","correspondingAuthor":false,"prefix":"","firstName":"Jingjing","middleName":"","lastName":"Guo","suffix":""},{"id":337863012,"identity":"74e56d56-2555-443e-b1e5-1b68a3d3f73d","order_by":2,"name":"Feng Xu","email":"","orcid":"","institution":"Taizhou Vocational and Technical College, Chemical Pharmaceutical Research Institute","correspondingAuthor":false,"prefix":"","firstName":"Feng","middleName":"","lastName":"Xu","suffix":""},{"id":337863013,"identity":"e8d4753e-95e2-4b3c-ad92-2b4a82f4e2c4","order_by":3,"name":"Mingjuan Ji","email":"","orcid":"","institution":"Zhejiang University of Technology","correspondingAuthor":false,"prefix":"","firstName":"Mingjuan","middleName":"","lastName":"Ji","suffix":""}],"badges":[],"createdAt":"2024-07-16 04:00:44","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4746837/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4746837/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":62267027,"identity":"de0f437f-ae7d-4586-84e8-5f710fc1eda6","added_by":"auto","created_at":"2024-08-12 09:34:29","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1923843,"visible":true,"origin":"","legend":"\u003cp\u003eStructure of PARP-1. (A) Schematic representation of human PARP-1 domains. (B) The crystal structure of the catalytic domain (PDB code 7kk5[6]).\u003c/p\u003e","description":"","filename":"image1.png","url":"https://assets-eu.researchsquare.com/files/rs-4746837/v1/ee90d1033756c0648ffc0c0a.png"},{"id":62267025,"identity":"65f3b1ba-f462-48fd-a3b9-711951b95dd2","added_by":"auto","created_at":"2024-08-12 09:34:29","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":54792,"visible":true,"origin":"","legend":"\u003cp\u003eThe structure of reported PARP-1 inhibitors\u003c/p\u003e","description":"","filename":"image2.png","url":"https://assets-eu.researchsquare.com/files/rs-4746837/v1/17fda1100b53b9e094de6ac8.png"},{"id":62267746,"identity":"0d7979c6-22a7-4c4f-95e9-805fdd31a93f","added_by":"auto","created_at":"2024-08-12 09:42:29","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":5202840,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eComprehensive Structural Analysis of Proteins. \u003c/strong\u003e(A) 4ZZZ X-Ray Diffraction Structure at 1.90 Å Resolution, 7KK2X-Ray Diffraction Structure at 1.70 Å Resolution, 7KK4 X-Ray Diffraction Structure at 1.96 Å Resolution, and 7KK5 X-Ray Diffraction Structure at 1.70 Å Resolution. (B) Distribution Analysis of Amino Acid Residues in Preferred Conformational Regions for 4ZZZ, 7KK2, 7KK4, and 7KK5. (C) Ramachandran Plots Highlighting Residue Distribution of 4ZZZ, 7KK2, 7KK4, and 7KK5. (D) Quality Assessment of Protein Models via ERRAT and Resolution.\u003c/p\u003e","description":"","filename":"image3.png","url":"https://assets-eu.researchsquare.com/files/rs-4746837/v1/ee77fa51d2d4818b7f77bea1.png"},{"id":62267028,"identity":"efc789e8-499b-4c94-be58-8dfc4d2ad759","added_by":"auto","created_at":"2024-08-12 09:34:29","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":1439197,"visible":true,"origin":"","legend":"\u003cp\u003eThe Hybrid Virtual screening workflow. (A) Overview of the virtual screening process targeting PARP-1. (B) Structural representation of the final selected 10 small molecules.\u003c/p\u003e","description":"","filename":"image4.png","url":"https://assets-eu.researchsquare.com/files/rs-4746837/v1/373fda91e964beb93e642bb9.png"},{"id":62267031,"identity":"21e69dcf-e801-481f-a350-627dbe8ce38f","added_by":"auto","created_at":"2024-08-12 09:34:29","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":6447992,"visible":true,"origin":"","legend":"\u003cp\u003eComprehensive Analysis of compounds 1, 3 ,6, 9 and PARP-1 Interactions. (A) RMSD values for PARP-1. (B) RMSD values for compounds 1, 3 ,6, 9. (C) RMSF of the backbones. (D) Gibbs Free Energy landscape of compound 1/PARP-1. (E) Three-dimensional visualization of the PARP-1-compound 1 interaction. (F) Two-dimensional representation of the PARP-1-compound 1 interaction. (G) Gibbs Free Energy landscape of compound 3/PARP-1.(H)Three-dimensional visualization of the compound 3/PARP-1 interaction. (I) Two-dimensional representation of the compound 3/PARP-1 interaction. (J) Gibbs Free Energy landscape of compound 6/PARP-1. (K) Three-dimensional visualization of the compound 6/PARP-1 interaction. (L) Two-dimensional representation of the compound 6/PARP-1 interaction. (M) Gibbs Free Energy landscape of compound 9/PARP-1. (N) Three-dimensional visualization of the compound 9/PARP-1 interaction. (O) Two-dimensional representation of the compound 9/PARP-1 interaction.\u003c/p\u003e","description":"","filename":"image5.png","url":"https://assets-eu.researchsquare.com/files/rs-4746837/v1/cd66e1ad0365f6b7ed970c4e.png"},{"id":62267030,"identity":"b3c64739-149e-4f0a-a307-7c5c34091fa5","added_by":"auto","created_at":"2024-08-12 09:34:29","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":608136,"visible":true,"origin":"","legend":"\u003cp\u003eThe major contributions of individual amino acid residues to the interaction complex of PARP1 with compounds (A) 1, (B) 3, (C) 6, (D) 9.\u003c/p\u003e","description":"","filename":"image6.png","url":"https://assets-eu.researchsquare.com/files/rs-4746837/v1/eb707f2ddc2ba077ea0c905a.png"},{"id":72777165,"identity":"3b6c149e-03af-4742-866a-f7121d562695","added_by":"auto","created_at":"2025-01-02 05:01:54","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":21261128,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4746837/v1/282eaf17-ced0-4d5d-a305-f14f54da2478.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Integrated Computational Pipeline for the Identification of Novel PARP-1 Inhibitors: Hybrid Virtual Screening and Molecular Dynamics Simulations","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003ePoly(ADP-ribose) polymerase 1 (PARP-1) is a versatile enzyme pivotal in post-translational modifications, playing a key role in several biological functions, such as DNA repair[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. PARP-1 is a prominent member of the PARP family, which encompasses 17 distinct members[\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. As depicted in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA, the structure of PARP-1 is comprising multiple domains: the \u003cem\u003eN\u003c/em\u003e-terminal zinc finger domains (Zn1, Zn2, and Zn3), the BRCT domain, the WGR domain, and the \u003cem\u003eC\u003c/em\u003e-terminal catalytic (CAT) domain[\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. The zinc finger domains Zn1 and Zn2, located at the \u003cem\u003eN\u003c/em\u003e-terminus, are specialized in recognizing specific DNA structures. The BRCT domain primarily facilitates automodification[\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. The CAT domain, which includes the HD and ART subdomains, is crucial for the enzyme's activity, mediating the addition of ADP-ribose polymers to target proteins, thereby enhancing the DNA repair mechanism as illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB[\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e].\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eDuring cell growth, DNA damage inevitably occurs due to internal and external factors[\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. Such damage can lead to cell cycle arrest or genomic instability, both of which are key characteristics that contribute to cellular transformation into cancer. PARP-1 inhibitors, a class of small molecule drugs targeting PARP-1, act through two primary mechanisms: synthetic lethality and immune activation. Synthetic lethality exploits DNA repair deficiencies in tumor cells. For instance, certain cancer cells, such as those with BRCA mutations found in breast and ovarian cancers, lack the homologous recombination repair (HRR) pathway, which is necessary for repairing double-strand breaks (DSBs). While normal cells repair DSBs through HRR, cancer cells are unable to do so[\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. As a result, PARP-1 inhibitors block the repair of single-strand breaks (SSBs), causing these breaks to evolve into DSBs during DNA replication, leading to the death of tumor cells while sparing normal cells. On the other hand, immune activation by PARP-1 inhibitors induces DNA damage that results in tumor cells producing more mutations and neoantigens, allowing them to be recognized and attacked by the immune system, thus achieving anticancer effects[\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. The first-generation PARP-1 inhibitors, based on the nicotinamide structure, incorporated electron-donating groups or bioisosteres to develop analogs like 3-aminobenzamide (3-AB), enhancing intermolecular interactions and inhibition efficacy. Second-generation PARP-1 inhibitors have improved upon drug design and structural changes to enhance potency and therapeutic outcomes. As shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e, since the market launch of the first PARP-1 inhibitor, Olaparib, in 2014, there have been significant breakthroughs in the development of PARP-1 inhibitors. Subsequently, the FDA has approved several others, including Niraparib, Rucaparib, and Talazoparib, highlighting the critical role of PARP-1 inhibitors in cancer treatment. Currently, numerous PARP-1 inhibitors are undergoing clinical trials. For instance, Simmiparib, developed by the Shanghai Institute of Pharmaceutical Research, substitutes the triazine ring in the structure of Olaparib with a triazole ring, achieving an inhibition rate five times higher than Olaparib with an IC\u003csub\u003e50\u003c/sub\u003e of 0.74 nM[\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. Sun et al. have synthesized derivatives containing phthalazin-1(2H)-one, where YCH1899 showed significant antiproliferative activity against Olaparib and Talazoparib-resistant cells with IC\u003csub\u003e50\u003c/sub\u003e values of 0.89 and 1.13 nM, respectively[\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e]. Despite the promising anticancer activity of PARP-1 inhibitors, they still exhibit side effects that limit their clinical applications. Therefore, it remains essential to explore new structural types of PARP-1 inhibitors to enhance efficacy and reduce side effects, thereby improving their clinical applicability and expanding their therapeutic potential.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eIn conventional computationally assisted drug design (CADD), the screening framework depends on the structures of proteins or ligands, with two main types: ligand-based (LBVS) and structure-based (SBVS) screening. LBVS concentrates on the structure of ligands but may neglect critical details of the target proteins, which can result in suboptimal binding effectiveness. On the other hand, SBVS uses protein structures but is often hampered by their limited availability and the difficulties in managing large-scale databases, which can lead to increased time and computational demands. Additionally, CADD faces challenges in precisely predicting how compounds will interact with target proteins. To overcome these limitations, researchers are increasingly incorporating artificial intelligence (AI) into CADD processes to boost both the accuracy and efficiency of predictions. AI technologies, especially deep learning and attention mechanisms, are adept at capturing and processing complex molecular features and interactions. This enhances the identification of active molecules and optimizes the screening workflow. For example, the Deep Docking platform employs iterative methods to improve the precision of predictions and increase the overall efficiency of the screening process, showcasing how AI can significantly expedite the drug discovery process by refining and accelerating CADD operations.\u003c/p\u003e \u003cp\u003eIn this study, we used hybrid virtual screening workflow to identified novel scaffold PAPR-1inhibitor molecules. The process involves multiple stages, including for Transfoxmol, KarmaDock, PLANET, and Vina for generated Conformation. We also applied MM/PBSA to evaluate further binding potential between candidate compounds and PARP-1. After several iterations of the workflow, our approach achieved promising results, with 10 compounds standing out. Furthermore, we selected three candidate molecules and performed a detailed study of the binding interactions between the compounds and their targets using molecular dynamics simulations. Therefore, we plan to conduct further molecular structure modification and optimization in the laboratory.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e"},{"header":"2. Computational details and methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1. Choice of PARP-1 structure and preparation\u003c/h2\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eIn our research, we conducted a thorough exploration of the UniProt database[\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e], specifically targeting the PARP-1_Human catalytic domain (662\u0026ndash;1011) with the UniProt ID P09874. From this inquiry, we obtained 55 X-ray co-crystal structures from the RCSB Protein Data Bank[\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. These structures were subjected to detailed analysis using the Structural Analysis and Verification Server (SAVES) v6.0, incorporating both the PROCHECK[\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e] and ERRAT[\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e] modules for validation. SAVES, which can be accessed at \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://saves.mbi.ucla.edu/\u003c/span\u003e\u003cspan address=\"https://saves.mbi.ucla.edu/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (as of January 13, 2024), provides a comprehensive suite for assessing the structural integrity and reliability of protein models. For the virtual screening, we specifically prepared the 7KK5 structure. Utilizing PyMOL[\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e], we meticulously removed water molecules and added hydrogen atoms to the structure.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2. Database preparation\u003c/h2\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eIn our study, we developed a virtual screening workflow centered on the Topscience database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.tsbiochem.com/\u003c/span\u003e\u003cspan address=\"https://www.tsbiochem.com/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), which comprises approximately 13\u0026nbsp;million molecules. The integrity and quality of our database are critical for the success of the screening process. To achieve this, we employed the RDKit[\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e] to implement a rigorous data preprocessing protocol. Our approach began with the identification and removal of duplicate molecular structures, ensuring that each molecule in our database is unique. This was followed by the parsing of each molecule, represented as a SMILES string, into a molecular structure object using RDKit. We discarded entries with invalid SMILES strings, indicating parsing failures. Subsequent steps involved the removal of salt components from molecules with RDKit's SaltRemover and the neutralization of molecular charges using the Uncharger function. These steps are crucial for isolating the principal active component of each molecule, making the dataset more uniform and facilitating consistent comparisons and analyses. Additionally, we conducted a boron valence verification to exclude molecules with boron atoms exhibiting explicit valences greater than three, as these do not conform to standard chemical validity. Finally, we standardized the molecular structures by converting them into the universally recognized SMILES format, preparing them for subsequent computational analyses.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e2.3. Virtual screening\u003c/h2\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003e \u003cb\u003eTransfoxmol\u003c/b\u003e \u003c/p\u003e \u003cp\u003eTransFoxMol (available on \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://github.com/gaojianl/TransFoxMol\u003c/span\u003e\u003cspan address=\"https://github.com/gaojianl/TransFoxMol\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e)[\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e] was conceived to augment the artificial intelligence's understanding of the intricate relationships between molecular structures and their properties. This method is notable for integrating a multi-scale 2D molecular environment into a cohesive framework that combines a graph neural network with a Transformer architecture. It uniquely utilizes pre-existing chemical maps to fine-tune its attention mechanism, thereby achieving a level of precision and efficiency beyond what current methodologies offer. The foundation for training this model was a dataset sourced from the ChEMBL database[\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e] as of January 12, 2024. This dataset was meticulously processed to exclude duplicate entries and those lacking essential information such as labels or SMILES notation. It was further enhanced by converting original IC\u003csub\u003e50\u003c/sub\u003e values to pIC\u003csub\u003e50\u003c/sub\u003e values, thus providing a more accurate reflection of molecular activity, represented as -lnIC\u003csub\u003e50\u003c/sub\u003e\u0026thinsp;=\u0026thinsp;pIC\u003csub\u003e50\u003c/sub\u003e.\u003c/p\u003e \u003cp\u003eThe experimental process commenced with the preparation of the PARP-1 dataset for regression analysis using the molnetdata.py script, followed by hyperparameter optimization executed through the \u0026lsquo;run.py search\u0026rsquo;. The model's training was precisely outlined with parameters set for regression analysis on the PARP-1 dataset (--task reg), utilizing a CUDA device, and specified settings including a batch size of 32, a total of 50 training epochs, a learning rate of 0.0005, both validation and test rates at 0.1, a seed value of 426, conducted over three folds. Additionally, the model featured a dropout rate of 8.85, eight attention heads, two attention layers, an output dimensionality of 128, and incorporated four distance thresholds (--D 4). The primary metric for assessing the model's performance was the Root Mean Square Error (RMSE), with the best validation RMSE recorded at 0.7569 and the best test RMSE at 0.8109. The testing phase reported a loss of 0.0974, a mean absolute error (MAE) of 0.4337, an RMSE of 0.5971, and an R\u003csup\u003e2\u003c/sup\u003e score of 0.7861. For the prediction stage, the model, trained on the PARP-1 dataset and encapsulated as PARP-1reg_44_0.7569000124931335.pkl, was applied to the prepared Topscience database to facilitate further molecular activity predictions.\u003c/p\u003e \u003cp\u003e \u003cb\u003eKarmaDock\u003c/b\u003e \u003c/p\u003e \u003cp\u003eWe utilized KarmaDock, a cutting-edge deep learning framework designed for ligand docking, which is publicly available on GitHub (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://github.com/schrojunzhang/KarmaDock\u003c/span\u003e\u003cspan address=\"https://github.com/schrojunzhang/KarmaDock\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e)[\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. This method encapsulates functionalities for expedited docking, generation and refinement of binding poses, and evaluation of binding affinities. KarmaDock's architecture is structured into a three-stage model: (1) It employs encoders to capture the protein and ligand's intramolecular interaction features. (2) It utilizes E(n) equivariant graph neural networks complemented with self-attention mechanisms to refine the ligand's pose, taking into account both protein-ligand and intraligand interactions. This stage is augmented by a post-processing step to guarantee the chemical validity of the generated structures. (3) It incorporates a mixture density network to quantitatively assess the binding affinity. In our experimental setup, the ligand was represented in '.smi' format for input purposes. The target protein, identified by PDB ID: 7kk5, along with its crystal-bound ligand, was retrieved from the RCSB Protein Data Bank[\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. Docking simulations were executed utilizing the virtual_screening_pipeline.py script.\u003c/p\u003e \u003cp\u003e \u003cb\u003ePLANET\u003c/b\u003e \u003c/p\u003e \u003cp\u003eThe PLANET model[\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e], accessible via \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://github.com/ComputArtCMCG/PLANET/\u003c/span\u003e\u003cspan address=\"https://github.com/ComputArtCMCG/PLANET/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e, was utilized, incorporating the 3D structure of a target protein's binding pocket represented as a graph and the 2D chemical structure of the ligand in \u0026lsquo;.sdf\u0026rsquo; format as its primary inputs. The model underwent an extensive training regimen following a tri-objective approach, aimed at assessing protein-ligand binding affinity, mapping out protein-ligand interaction sites, and creating the ligand distance matrix. For our experiments, the ligand's chemical structure, provided in \u0026lsquo;.sdf\u0026rsquo; format, served as the input. The protein of interest, bearing the PDB ID: 7kk5, and its crystal-bound ligand were sourced from the RCSB database[\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. The docking analysis was performed using the PLANET_run.py script.\u003c/p\u003e \u003cp\u003e \u003cb\u003eAutoDock Vina\u003c/b\u003e \u003c/p\u003e \u003cp\u003eAutoDock Vina (version 1.2.3)[\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e, \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e] was deployed for ligand preparation, utilizing prepare_ligand4.py, and for generating the configuration file through prepare_pdb_split_alt_confs.py, which led to the creation of configure.txt. This file specified parameters such as: center_x = -36.91, center_y\u0026thinsp;=\u0026thinsp;5.89, center_z = -7.45; size_x\u0026thinsp;=\u0026thinsp;30, size_y\u0026thinsp;=\u0026thinsp;30, size_z\u0026thinsp;=\u0026thinsp;30; energy_range\u0026thinsp;=\u0026thinsp;3; exhaustiveness\u0026thinsp;=\u0026thinsp;48; num_modes\u0026thinsp;=\u0026thinsp;20. Following this, prepare_receptor4.py was applied for receptor preparation, which included the removal of water molecules and the confirmation of hydrogen atoms' presence. The docking procedure was subsequently executed using Vina.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eFiltered and Cluster\u003c/b\u003e \u003c/p\u003e \u003cp\u003eIn the final stage of virtual screening, our methodology incorporated a suite of Python libraries for data manipulation and analysis, including RDKit for cheminformatics, Matplotlib[\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e] and Pandas[\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e] for data visualization and manipulation, NumPy[\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e] for numerical operations, SciPy[\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e] for distance calculations, and sklearn for clustering algorithms. Our process commenced with loading a dataset of molecular docking scores, derived from both Vina and PLANET analyses, from a CSV file. This dataset was then refined to isolate molecules exhibiting PLANET scores above 10 and Vina scores below \u0026minus;\u0026thinsp;10, resulting in a cohort of 191 molecules targeted for clustering. The clustering phase involved generating Morgan fingerprints[\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e] for each molecule. These fingerprints are structured as fixed-length bit vectors, encoding the presence or absence of particular molecular substructures, thereby serving as a compact representation of molecular features. Subsequent steps entailed calculating pairwise Jaccard distances among these fingerprints to assess chemical dissimilarities between molecules. The Jaccard distance[\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e], a metric of dissimilarity between two sets, facilitated the quantitative analysis of chemical diversity within our dataset. We then applied agglomerative hierarchical clustering to this distance matrix. This clustering technique progressively merges pairs of clusters based on their proximity, utilizing precomputed Jaccard distances as the affinity metric and adopting an average linkage strategy to gauge distances between clusters. Through this method, each molecule was assigned to one of ten predetermined clusters, reflecting their chemical similarities.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e2.4. Molecular dynamics simulation\u003c/h2\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eIn our research, we utilized molecular dynamics simulations to investigate protein-ligand interactions, employing GROMACS[\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e] (version 2020.7) and structures from AutoDock Vina[\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e]. We formatted protein and ligand structures in PDB, refining proteins with pdbfixer (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://github.com/openmm/pdbfixer\u003c/span\u003e\u003cspan address=\"https://github.com/openmm/pdbfixer\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) and using the amber99sb-ildn force field. Our simulations took place in a 1.5 nm cubic box using the tip3p water model. For ligand-containing systems, we derived parameters for both ligands and receptors, converting ligand structures to mol2 format with Open Babel, and parameterizing with ACPYPE (version 1.3.0)[\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e]. We created GROMACS-compatible receptor topology files with pdb2gmx and integrated them with ligand topologies into a unified file. We applied the -DFLEXIBLE option before starting energy minimization with the Conjugate Gradient algorithm, interspersed with steepest descent corrections to prevent local minima, ceasing minimization when forces dropped below 1000 kJ/mol/nm. For the NVT equilibration phase, we used a leap-frog integrator over 500 ps with data recording at 1.0 ps intervals. Spatial searches utilized the grid method, with long-range electrostatics managed via the Particle Mesh Ewald method. Temperature was controlled using the Berendsen thermostat at 300 K. The NPT phase followed with the same settings, and pressure controlled by the Berendsen method at 1.0 bar isotropic coupling. Our production phase spanned 150 ns, capturing data every 100 ps, with all hydrogen bonds constrained by the LINCS algorithm and thermodynamic parameters managed using the V-rescale and Parrinello-Rahman methods.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003e2.5. MM/PBSA\u003c/h2\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eWe conducted MM-PBSA calculations by employing a custom bash script designed specifically for processing GROMACS[\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e] (version 2020.7) molecular dynamics simulations and Adaptive Poisson-Boltzmann Solver (APBS) (version 3.0.0)[\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e] for calculating electrostatic potentials. This script facilitated the efficient preparation of input files by automating the extraction of trajectory data (.xtc), topology (.tpr), and index files (*.ndx). The resources were made available through Jerkwin's gmxtools repository on GitHub (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://github.com/Jerkwin/gmxtools\u003c/span\u003e\u003cspan address=\"https://github.com/Jerkwin/gmxtools\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) [\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]. The streamlined procedure included preparing input files from GROMACS outputs, setting up APBS[\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e] calculations, and estimating binding free energies. For the segregation of ligands and proteins, specific index groups were identified. In the configuration file (configure.txt), we meticulously defined parameters, including those for electrostatic and van der Waals interactions, the Debye-H\u0026uuml;ckel screening method for MM (Coulombic) energies, and criteria for selecting residues for the calculations[\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e]. These parameters were crucial for ensuring the accuracy of simulations, encompassing solvent models and boundary conditions, thus facilitating precise electrostatic potential evaluation by APBS. Energy calculations, encompassing molecular mechanics, polar, and non-polar solvation energies, were performed, leading to the estimation of binding free energies. The analysis was based on trajectory data (.xtc), topology information (.tpr), and index files (*.ndx) derived from the final 50 ns of 150 ns molecular dynamics simulation.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"3. Results and discussion","content":"\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003e3.1. The evaluation of PARP-1 structures\u003c/h2\u003e \u003cp\u003eIn our quest to identify the optimal PARP-1 protein structure for our study, we initiated our search within the UniProt database[\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e] for the PARP-1 enzyme specific to Homo sapiens (Human), bearing the UniProt ID P09874 (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.uniprot.org\u003c/span\u003e\u003cspan address=\"https://www.uniprot.org\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e, access date: 2024-1-13), and identified 55 PARP1_HUMAN structures containing catalytic domain (662\u0026ndash;1011). Our selection criteria excluded any structures exhibiting mutations or those with a resolution exceeding 2\u0026Aring;, to ensure high fidelity in our computational analyses (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA, Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eFurther scrutiny was directed towards four remaining structures, evaluated using the SAVES v6.0 platform, a comprehensive tool for assessing structural integrity and reliability. This analysis encompassed several aspects; firstly, VERIFY 3D inspection was applied to assess the compatibility of each structure's amino acid sequence with its 3D structure, with all nine structures surpassing the threshold of 80 points (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB, Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Subsequent examination employed the PROCHECK module, which gauges the stereochemical quality of the protein structures. Among the structures analyzed, 7KK4 was noted for having less than 90% of its residues in the most preferred regions, leading to its exclusion from further consideration (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eC, Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eAdditionally, the ERRAT module, designed for the evaluation of crystallography-derived protein structures, provided an overall quality factor for each structure (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eD, Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Notably, structure 7KK2 emerged as the optimal choice based on this metric; however, it was ultimately deemed unsuitable due to the absence of small molecule compounds within its structure. Considering both resolution and PROCHECK scores, we concluded that structure 7KK5 represented the best candidate for the foundation of our virtual screening efforts in this study.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003ePARP1 structure Quality Metrics and Ramachandran Plot Statistics\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"7\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\"\u0026minus;\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePDB\u003c/p\u003e \u003cp\u003eID\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMutation\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eResolution\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eReleased\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eOverall quality factor\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eVERIFY\u003c/p\u003e \u003cp\u003e3D\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eRamachandran\u003c/p\u003e \u003cp\u003eplot\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e4ZZZ\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.90 \u0026Aring;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026minus;\" colname=\"c4\"\u003e \u003cp\u003e2015-08-12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e95.0966\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003ePass\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e91.9% core\u003c/p\u003e \u003cp\u003e8.1% allow\u003c/p\u003e \u003cp\u003e0.0% gener\u003c/p\u003e \u003cp\u003e0.0% disal\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e7KK2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.70 \u0026Aring;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026minus;\" colname=\"c4\"\u003e \u003cp\u003e2021-01-06\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e97.0458\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003ePass\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e92.4% core\u003c/p\u003e \u003cp\u003e7.6% allow\u003c/p\u003e \u003cp\u003e0.0% gener\u003c/p\u003e \u003cp\u003e0.0% disa\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e7KK4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.96 \u0026Aring;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026minus;\" colname=\"c4\"\u003e \u003cp\u003e2021-01-06\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e95.3632\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003ePass\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e89.8% core\u003c/p\u003e \u003cp\u003e10.0% allow\u003c/p\u003e \u003cp\u003e0.2% gener\u003c/p\u003e \u003cp\u003e0.0% disal\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e7KK5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eNo\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.70 \u0026Aring;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\"\u0026minus;\" colname=\"c4\"\u003e \u003cp\u003e2021-01-06\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e96.5839\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003ePass\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e92.2% core\u003c/p\u003e \u003cp\u003e7.8% allow\u003c/p\u003e \u003cp\u003e0.0% gener\u003c/p\u003e \u003cp\u003e0.0% disal\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003e3.2. Hybrid Virtual screening\u003c/h2\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eIn our pursuit to identify potent PARP-1 inhibitors, we adopted an integrative approach, combining AI-driven screening methodologies with conventional computer-aided drug design (CADD) tools (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA). Initially, we leveraged TransFoxMol, a cutting-edge tool recognized for its innovative integration of multi-scale 2D molecular environments within a graph neural network and Transformer architecture framework. TransFoxMol excels in correlating molecular structures with their properties, thereby facilitating an efficient initial screening of the PAPR1 activity dataset sourced from the ChEMBL database. Through this initial phase, we utilized TransFoxMol to screen the extensive Topscience database (total of 13\u0026nbsp;million molecules), filtering for molecules predicted with a score above 7.5, resulting in 189,480 candidates.\u003c/p\u003e \u003cp\u003eFor the subsequent round of screening, we employed KarmaDock, an AI-based docking tool renowned for its precision in rapidly predicting the binding strength of protein-ligands. By setting a threshold Karma score of above 80, we further refined our selection to 3,471 molecules. The third phase of our screening process involved the use of PLANET, another AI-based tool designed to predict protein-ligand binding affinity with high accuracy. Despite the strengths of AI in predicting molecular activity and affinity, these tools often fall short in accurately generating 3D conformations of compounds. To address this limitation, we incorporated AutoDock Vina, a traditional molecular docking tool, to enhance the precision of compound conformations. Focusing on molecules that achieved a PLANET score above 10 and a Vina score below \u0026minus;\u0026thinsp;10 kcal/mol, we distilled our list to 192 compounds.\u003c/p\u003e \u003cp\u003eTo assess the structural diversity within our dataset, we utilized Morgan molecular fingerprints, a method allowing for the detailed categorization of molecules into one of ten pre-established clusters reflecting their chemical similarity. This clustering enabled us to distill our findings further to ten unique molecules (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB and Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). The selection criterion for further analysis was primarily based on achieving the highest Vina docking scores, from which compounds 1, 3, 6, and 9 were chosen for advanced Molecular Dynamics (MD) simulations.\u003c/p\u003e \u003cp\u003eThis approach, a synergy of hybrid virtual screening and structural clustering, integrates the predictive prowess of artificial intelligence with the robust methodologies of CADD technologies. By doing so, we not only streamline the identification process of promising PARP-1 inhibitor candidates but also furnish future researchers with novel PARP-1 scaffolds.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eEach stage score of the final selected 10 small molecules.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"7\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEntry\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eInChIKey\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eTransfoxmol\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eKarmaDock\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003ePLANET\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eVina\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eMM/PBSA\u003c/p\u003e \u003cp\u003e(kcal/mol)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCompd. 1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eInChIKey\u0026thinsp;=\u0026thinsp;ORBZKIMBPNECHS-UHFFFAOYSA-N\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e7.715\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e93.262\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e16.094\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e12.887\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e-24.013\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCompd. 2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eInChIKey\u0026thinsp;=\u0026thinsp;YMJJZAYPAWMBMO-PTTDRDKLNA-N\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e7.842\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e82.217\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e11.283\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e10.796\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCompd. 3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eInChIKey\u0026thinsp;=\u0026thinsp;ULUVBBCDTWTZLD-UHFFFAOYSA-N\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e8.541\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e94.830\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e12.847\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e14.211\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e-40.081\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCompd. 4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eInChIKey\u0026thinsp;=\u0026thinsp;HYUHVOLQYDHZHQ-UHFFFAOYSA-N\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e7.587\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e85.068\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e10.942\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e11.482\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCompd. 5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eInChIKey\u0026thinsp;=\u0026thinsp;DJIKKGWKXJGKRI-WYOOIXGGSA-N\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e7.503\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e108.382\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e15.740\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e10.798\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCompd. 6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eInChIKey\u0026thinsp;=\u0026thinsp;DSYYGKZYBMLNEM-FLFKKZLDSA-N\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e7.778\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e88.667\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e11.766\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e11.573\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e-20.806\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCompd.7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eInChIKey\u0026thinsp;=\u0026thinsp;NDIZSXKXYRQMAU-SXGWCWSVSA-N\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e7.640\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e81.205\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e12.289\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e10.836\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCompd. 8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eInChIKey\u0026thinsp;=\u0026thinsp;ANAHWPGZGLJKPD-UHFFFAOYNA-N\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e7.589\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e80.588\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e10.901\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e10.796\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCompd. 9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eInChIKey\u0026thinsp;=\u0026thinsp;RVNGBBQZELJWBB-UHFFFAOYSA-N\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e8.038\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e93.073\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e12.382\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e11.851\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e-12.880\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCompd.10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eInChIKey\u0026thinsp;=\u0026thinsp;GCVOEPYFYISOOG-UHFFFAOYSA-N\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e7.597\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e85.930\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e11.910\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e11.491\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e-\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003e3.3. Determination of structural stability and dynamical of the complexes\u003c/h2\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eThe determination of structural stability and dynamics of complexes is a critical phase in the computational analysis of protein-ligand interactions. This step involves conducting Molecular Dynamics (MD) simulations, a computational method that enables the exploration of the physical movements of atoms and molecules over time. Here, our objective is to evaluate the behavior of selected PARP-1 inhibitor complexes under simulated physiological conditions, providing insights into their conformational stability, flexibility, and overall dynamic behavior. As depicted in Figs.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA and B, the trajectories of backbones and ligands reach equilibrium in the final 10 ns. Figure\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eC illustrates the RMSF plot for PARP-1, demonstrating amino acid residue fluctuations ranging from 0.1 \u0026Aring; to 0.6 \u0026Aring;.\u003c/p\u003e \u003cp\u003eFurthermore, an analysis of the Gibbs Free Energy landscape (Figs.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eD, G, J, M) identified the most energetically favorable compounds 1, 3, 6, 9, and PARP-1. Notably, compound 6 forms multiple hydrogen bond interactions with Arg878, Gly863, and Ser904 of PARP-1, while compound 6 interacts via a hydrogen bond with Arg865 of PARP-1. Additionally, compound 9 forms multiple hydrogen bond interactions with Arg878, Asp770, His862, Ser864, Gly863, Ser904, Asp766, and Met980 of PARP-1.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003e3.4. Binding free energy calculations\u003c/h2\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003eIn the realm of drug discovery, accurately assessing the binding free energy is essential for evaluating the efficacy and specificity of drug candidates. The Molecular Mechanics/Poisson-Boltzmann Surface Area (MM/PBSA) method stands out as a crucial technique in this process due to its optimal balance between computational intensity and accuracy. In our study, MM/PBSA analyses were performed using data from molecular dynamics (MD) simulation trajectories. These analyses shed light on the contributions of individual amino acid residues to the overall binding free energy in various PARP-1/compound complexes. As depicted in Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e, negative values indicate beneficial interactions that enhance the stability of the complex. For instance, the binding free energy of compound 1 with PARP-1 was calculated to be -24.013 kcal/mol, with significant contributions from residues like PRO881, TYR907, and TYR889 (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eA, Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). Similarly, compound 3 demonstrated a binding free energy of -40.081 kcal/mol with notable contributions from TYR896, TYR907, and GLU763 (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eB, Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). Compound 6 showed a binding free energy of -20.806 kcal/mol, influenced significantly by TYR896, TYR907, and GLU889 (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eC, Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). Lastly, a repeat analysis for compound 6/PARP-1 presented a binding free energy of -12.880 kcal/mol with contributions from TYR907, TYR896, and ASN868 being particularly impactful (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eD, Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"4. Conclusions","content":"\u003cp\u003eDespite the promising anticancer activity of PARP-1 inhibitors, they still exhibit side effects that limit their clinical applications. Therefore, exploring new structural types of PARP-1 inhibitors to enhance efficacy and reduce side effects remains essential, thereby improving their clinical applicability and expanding their therapeutic potential. In this study, we conducted a comprehensive screening process aimed at identifying potential PARP-1 inhibitors. Our methodology integrated Transfoxmol, KarmaDock, and PLANET with AutoDock Vina, utilizing it for rescoring and generating conformations. This combined approach successfully identified ten potent PARP-1 inhibitors from the Topscience database. In the final round of screening, we employed clustering to obtain ten structurally diverse compounds, providing insights for future PARP-1 inhibitor design. Furthermore, molecular dynamics simulations and MM/PBSA were performed to precisely elucidate the binding kinetics of compounds 1, 3, 6, and 9 with PARP-1, offering a clear understanding of their interaction mechanisms. An analysis of the Gibbs Free Energy landscape identified the most energetically favorable compounds 1, 3, 6, 9, and PARP-1. Notably, compound 6 formed multiple hydrogen bond interactions with Arg878, Gly863, and Ser904 of PARP-1, while compound 6 interacted via a hydrogen bond with Arg865 of PARP-1. Additionally, compound 9 formed multiple hydrogen bond interactions with Arg878, Asp770, His862, Ser864, Gly863, Ser904, Asp766, and Met980 of PARP-1.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgments\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work was supported by the [Taizhou Science and technology plan] under Grant [number\u0026nbsp;22ywa61]; [Zhejiang soft Science research program] under Grant [number\u0026nbsp;2024C35015]; and [Taizhou Vocational and Technical College doctoral research fund] under Grant [number\u0026nbsp;14071191].\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDisclosure statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThere are no conflicts of interest to declare.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIf you need research data related to the article, please request it from the corresponding author.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eRay Chaudhuri, A.; Nussenzweig, A. The Multifaceted Roles of PARP1 in DNA Repair and Chromatin Remodelling. Nat Rev Mol Cell Biol 2017, \u003cem\u003e18\u003c/em\u003e, 610\u0026ndash;621, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/nrm.2017.53\u003c/span\u003e\u003cspan address=\"10.1038/nrm.2017.53\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAlemasova, E.E.; Lavrik, O.I. Poly(ADP-Ribosyl)Ation by PARP1: Reaction Mechanism and Regulatory Proteins. Nucleic Acids Research 2019, \u003cem\u003e47\u003c/em\u003e, 3811\u0026ndash;3827, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/nar/gkz120\u003c/span\u003e\u003cspan address=\"10.1093/nar/gkz120\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLangelier, M.-F.; Pascal, J.M. PARP-1 Mechanism for Coupling DNA Damage Detection to Poly(ADP-Ribose) Synthesis. Current Opinion in Structural Biology 2013, \u003cem\u003e23\u003c/em\u003e, 134\u0026ndash;143, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.sbi.2013.01.003\u003c/span\u003e\u003cspan address=\"10.1016/j.sbi.2013.01.003\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFu, X.; Li, P.; Zhou, Q.; He, R.; Wang, G.; Zhu, S.; Bagheri, A.; Kupfer, G.; Pei, H.; Li, J. Mechanism of PARP Inhibitor Resistance and Potential Overcoming Strategies. Genes \u0026amp; Diseases 2024, \u003cem\u003e11\u003c/em\u003e, 306\u0026ndash;320, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.gendis.2023.02.014\u003c/span\u003e\u003cspan address=\"10.1016/j.gendis.2023.02.014\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKanev, P.-B.; Atemin, A.; Stoynov, S.; Aleksandrov, R. PARP1 Roles in DNA Repair and DNA Replication: The Basi(c)s of PARP Inhibitor Efficacy and Resistance. Seminars in Oncology 2024, \u003cem\u003e51\u003c/em\u003e, 2\u0026ndash;18, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1053/j.seminoncol.2023.08.001\u003c/span\u003e\u003cspan address=\"10.1053/j.seminoncol.2023.08.001\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRyan, K.; Bolanos, B.; Smith, M.; Palde, P.B.; Cuenca, P.D. Structure of the Catalytic Domain of PARP1 in Complex with Niraparib. 2021, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2210/pdb7KK5/pdb\u003c/span\u003e\u003cspan address=\"10.2210/pdb7KK5/pdb\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDHANYA K. NAMBIAR, 3, DEEPALI MISHRA, RANA P. SINGH Targeting DNA Repair for Cancer Treatment: Lessons from PARP Inhibitor Trials. Oncology Research 2023, \u003cem\u003e31\u003c/em\u003e, 405\u0026ndash;421, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.32604/or.2023.028310\u003c/span\u003e\u003cspan address=\"10.32604/or.2023.028310\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhao, Y.; Zhang, L.-X.; Jiang, T.; Long, J.; Ma, Z.-Y.; Lu, A.-P.; Cheng, Y.; Cao, D.-S. The Ups and Downs of Poly(ADP-Ribose) Polymerase-1 Inhibitors in Cancer Therapy\u0026ndash;Current Progress and Future Direction. European Journal of Medicinal Chemistry 2020, \u003cem\u003e203\u003c/em\u003e, 112570, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.ejmech.2020.112570\u003c/span\u003e\u003cspan address=\"10.1016/j.ejmech.2020.112570\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLord, C.J.; Tutt, A.N.J.; Ashworth, A. Synthetic Lethality and Cancer Therapy: Lessons Learned from the Development of PARP Inhibitors. Annual Review of Medicine 2015, \u003cem\u003e66\u003c/em\u003e, 455\u0026ndash;470.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAbbotts, R.; Wilson, D.M. Coordination of DNA Single Strand Break Repair. Free Radical Biology and Medicine 2017, \u003cem\u003e107\u003c/em\u003e, 228\u0026ndash;244, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.freeradbiomed.2016.11.039\u003c/span\u003e\u003cspan address=\"10.1016/j.freeradbiomed.2016.11.039\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYe, N.; Chen, C.-H.; Chen, T.; Song, Z.; He, J.-X.; Huan, X.-J.; Song, S.-S.; Liu, Q.; Chen, Y.; Ding, J.; et al. Design, Synthesis, and Biological Evaluation of a Series of Benzo[de][1,7]Naphthyridin-7(8H)-Ones Bearing a Functionalized Longer Chain Appendage as Novel PARP1 Inhibitors. J. Med. Chem. 2013, \u003cem\u003e56\u003c/em\u003e, 2885\u0026ndash;2903, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1021/jm301825t\u003c/span\u003e\u003cspan address=\"10.1021/jm301825t\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSun, Y.; Yang, H.; Yuan, J.; Wang, L.; Song, S.; Chen, R.; Bao, X.; Jia, L.; Yang, T.; Zhang, X.; et al. YCH1899, a Highly Effective Phthalazin-1(2H)-One Derivative That Overcomes Resistance to Prior PARP Inhibitors. J. Med. Chem. 2023, \u003cem\u003e66\u003c/em\u003e, 12284\u0026ndash;12303, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1021/acs.jmedchem.3c00821\u003c/span\u003e\u003cspan address=\"10.1021/acs.jmedchem.3c00821\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eThe UniProt Consortium UniProt: The Universal Protein Knowledgebase in 2023. Nucleic Acids Research 2023, \u003cem\u003e51\u003c/em\u003e, D523\u0026ndash;D531, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/nar/gkac1052\u003c/span\u003e\u003cspan address=\"10.1093/nar/gkac1052\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRCSB Protein Data Bank: Efficient Searching and Simultaneous Access to One Million Computed Structure Models Alongside the PDB Structures Enabled by Architectural Advances. Journal of Molecular Biology 2023, \u003cem\u003e435\u003c/em\u003e, 167994, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.jmb.2023.167994\u003c/span\u003e\u003cspan address=\"10.1016/j.jmb.2023.167994\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLaskowski, R.A.; Rullmannn, J.A.; MacArthur, M.W.; Kaptein, R.; Thornton, J.M. AQUA and PROCHECK-NMR: Programs for Checking the Quality of Protein Structures Solved by NMR. J Biomol NMR 1996, \u003cem\u003e8\u003c/em\u003e, 477\u0026ndash;486, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/BF00228148\u003c/span\u003e\u003cspan address=\"10.1007/BF00228148\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLaskowski, R.A.; MacArthur, M.W.; Moss, D.S.; Thornton, J.M. PROCHECK: A Program to Check the Stereochemical Quality of Protein Structures. J Appl Cryst 1993, \u003cem\u003e26\u003c/em\u003e, 283\u0026ndash;291, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1107/S0021889892009944\u003c/span\u003e\u003cspan address=\"10.1107/S0021889892009944\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eColovos, C.; Yeates, T.O. Verification of Protein Structures: Patterns of Nonbonded Atomic Interactions. Protein Sci 1993, \u003cem\u003e2\u003c/em\u003e, 1511\u0026ndash;1519, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/pro.5560020916\u003c/span\u003e\u003cspan address=\"10.1002/pro.5560020916\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchr\u0026ouml;dinger, LLC The PyMOL Molecular Graphics System, Version 1.8 2015.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLandrum, G. RDKit: A Software Suite for Cheminformatics, Computational Chemistry, and Predictive Modeling 2019.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGao, J.; Shen, Z.; Xie, Y.; Lu, J.; Lu, Y.; Chen, S.; Bian, Q.; Guo, Y.; Shen, L.; Wu, J.; et al. TransFoxMol: Predicting Molecular Property with Focused Attention. Briefings in Bioinformatics 2023, \u003cem\u003e24\u003c/em\u003e, bbad306, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/bib/bbad306\u003c/span\u003e\u003cspan address=\"10.1093/bib/bbad306\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDavies, M.; Nowotka, M.; Papadatos, G.; Dedman, N.; Gaulton, A.; Atkinson, F.; Bellis, L.; Overington, J.P. ChEMBL Web Services: Streamlining Access to Drug Discovery Data and Utilities. Nucleic Acids Res 2015, \u003cem\u003e43\u003c/em\u003e, W612\u0026ndash;W620, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/nar/gkv352\u003c/span\u003e\u003cspan address=\"10.1093/nar/gkv352\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGaulton, A.; Hersey, A.; Nowotka, M.; Bento, A.P.; Chambers, J.; Mendez, D.; Mutowo, P.; Atkinson, F.; Bellis, L.J.; Cibri\u0026aacute;n-Uhalte, E.; et al. The ChEMBL Database in 2017. Nucleic Acids Research 2017, \u003cem\u003e45\u003c/em\u003e, D945\u0026ndash;D954, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/nar/gkw1074\u003c/span\u003e\u003cspan address=\"10.1093/nar/gkw1074\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang, X.; Zhang, O.; Shen, C.; Qu, W.; Chen, S.; Cao, H.; Kang, Y.; Wang, Z.; Wang, E.; Zhang, J.; et al. Efficient and Accurate Large Library Ligand Docking with KarmaDock. Nat Comput Sci 2023, \u003cem\u003e3\u003c/em\u003e, 789\u0026ndash;804, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s43588-023-00511-5\u003c/span\u003e\u003cspan address=\"10.1038/s43588-023-00511-5\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang, X.; Gao, H.; Wang, H.; Chen, Z.; Zhang, Z.; Chen, X.; Li, Y.; Qi, Y.; Wang, R. PLANET: A Multi-Objective Graph Neural Network Model for Protein\u0026ndash;Ligand Binding Affinity Prediction. J. Chem. Inf. Model. 2023, \u003cdiv class=\"ExternalRefDOI\"\u003eacs.jcim.3c00253\u003c/div\u003e, doi:10.1021/acs.jcim.3c00253.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEberhardt, J.; Santos-Martins, D.; Tillack, A.F.; Forli, S. AutoDock Vina 1.2.0: New Docking Methods, Expanded Force Field, and Python Bindings. J. Chem. Inf. Model. 2021, \u003cem\u003e61\u003c/em\u003e, 3891\u0026ndash;3898, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1021/acs.jcim.1c00203\u003c/span\u003e\u003cspan address=\"10.1021/acs.jcim.1c00203\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTrott, O.; Olson, A.J. AutoDock Vina: Improving the Speed and Accuracy of Docking with a New Scoring Function, Efficient Optimization, and Multithreading. Journal of Computational Chemistry 2010, \u003cem\u003e31\u003c/em\u003e, 455\u0026ndash;461, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/jcc.21334\u003c/span\u003e\u003cspan address=\"10.1002/jcc.21334\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHunter, J.D. Matplotlib: A 2D Graphics Environment. Computing in Science \u0026amp; Engineering 2007, \u003cem\u003e9\u003c/em\u003e, 90\u0026ndash;95, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1109/MCSE.2007.55\u003c/span\u003e\u003cspan address=\"10.1109/MCSE.2007.55\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eteam, T. pandas development Pandas-Dev/Pandas: Pandas 2024.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHarris, C.R.; Millman, K.J.; Van Der Walt, S.J.; Gommers, R.; Virtanen, P.; Cournapeau, D.; Wieser, E.; Taylor, J.; Berg, S.; Smith, N.J.; et al. Array Programming with NumPy. Nature 2020, \u003cem\u003e585\u003c/em\u003e, 357\u0026ndash;362, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41586-020-2649-2\u003c/span\u003e\u003cspan address=\"10.1038/s41586-020-2649-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVirtanen, P.; Gommers, R.; Oliphant, T.E.; Haberland, M.; Reddy, T.; Cournapeau, D.; Burovski, E.; Peterson, P.; Weckesser, W.; Bright, J.; et al. SciPy 1.0: Fundamental Algorithms for Scientific Computing in Python. Nat Methods 2020, \u003cem\u003e17\u003c/em\u003e, 261\u0026ndash;272, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41592-019-0686-2\u003c/span\u003e\u003cspan address=\"10.1038/s41592-019-0686-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMorgan, H.L. The Generation of a Unique Machine Description for Chemical Structures-A Technique Developed at Chemical Abstracts Service. J. Chem. Doc. 1965, \u003cem\u003e5\u003c/em\u003e, 107\u0026ndash;113, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1021/c160017a018\u003c/span\u003e\u003cspan address=\"10.1021/c160017a018\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHancock, J.M. Jaccard Distance (Jaccard Index, Jaccard Similarity Coefficient). In \u003cem\u003eDictionary of Bioinformatics and Computational Biology\u003c/em\u003e; 2014 ISBN 978-0-471-65012-6.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAbraham, M.J.; Murtola, T.; Schulz, R.; P\u0026aacute;ll, S.; Smith, J.C.; Hess, B.; Lindahl, E. GROMACS: High Performance Molecular Simulations through Multi-Level Parallelism from Laptops to Supercomputers. SoftwareX 2015, \u003cem\u003e1\u003c/em\u003e, 19\u0026ndash;25, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.softx.2015.06.001\u003c/span\u003e\u003cspan address=\"10.1016/j.softx.2015.06.001\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTrott, O.; Olson, A. Software News and Update AutoDock Vina: Improving the Speed and Accuracy of Docking with a New Scoring Function, Efficient Optimization, and Multithreading. JOURNAL OF COMPUTATIONAL CHEMISTRY 2010, \u003cem\u003e31\u003c/em\u003e, 455\u0026ndash;461, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/jcc.21334\u003c/span\u003e\u003cspan address=\"10.1002/jcc.21334\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSousa da Silva, A.W.; Vranken, W.F. ACPYPE - AnteChamber PYthon Parser interfacE. BMC Research Notes 2012, \u003cem\u003e5\u003c/em\u003e, 367, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/1756-0500-5-367\u003c/span\u003e\u003cspan address=\"10.1186/1756-0500-5-367\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJurrus, E.; Engel, D.; Star, K.; Monson, K.; Brandi, J.; Felberg, L.E.; Brookes, D.H.; Wilson, L.; Chen, J.; Liles, K.; et al. Improvements to the APBS Biomolecular Solvation Software Suite. Protein Science 2018, \u003cem\u003e27\u003c/em\u003e, 112\u0026ndash;128, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1002/pro.3280\u003c/span\u003e\u003cspan address=\"10.1002/pro.3280\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi, J. Gmxtools 2022.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGenheden, S.; Ryde, U. The MM/PBSA and MM/GBSA Methods to Estimate Ligand-Binding Affinities. Expert Opinion on Drug Discovery 2015, \u003cem\u003e10\u003c/em\u003e, 449\u0026ndash;461, doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1517/17460441.2015.1032936\u003c/span\u003e\u003cspan address=\"10.1517/17460441.2015.1032936\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"PARP-1, Hybrid Virtual Screening, Molecular Dynamics","lastPublishedDoi":"10.21203/rs.3.rs-4746837/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4746837/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eDespite the promising anticancer properties of PARP-1 inhibitors, their clinical use is hindered by side effects. It is crucial to explore new structural variants of these inhibitors to increase efficacy and minimize side effects, enhancing their clinical viability and therapeutic scope. In this study, we developed a virtual screening workflow that synergistically integrates Transfoxmol, KarmaDock, and PLANET with AutoDock Vina's capabilities. Through structural clustering, we identified ten potential PARP-1 inhibitors. Additionally, through molecular dynamics simulations and MM/PBSA, we elucidated the binding modes of compounds 1, 3, 6, and 9 with PARP-1, providing insights for drug development.\u003c/p\u003e","manuscriptTitle":"Integrated Computational Pipeline for the Identification of Novel PARP-1 Inhibitors: Hybrid Virtual Screening and Molecular Dynamics Simulations","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-08-12 09:34:24","doi":"10.21203/rs.3.rs-4746837/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"87e6c762-300a-45f3-9f07-79231af51801","owner":[],"postedDate":"August 12th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":35792321,"name":"Biological sciences/Cancer"},{"id":35792322,"name":"Biological sciences/Drug discovery"},{"id":35792323,"name":"Physical sciences/Chemistry/Theoretical chemistry/Computational chemistry"},{"id":35792324,"name":"Physical sciences/Chemistry/Theoretical chemistry/Molecular dynamics"}],"tags":[],"updatedAt":"2025-01-02T04:53:35+00:00","versionOfRecord":[],"versionCreatedAt":"2024-08-12 09:34:24","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-4746837","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4746837","identity":"rs-4746837","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}