Active Learning-Driven Discovery and Dynamics Simulation of Novel SQLE Inhibitors with ADMET Analysis and Molecular Modification

Active Learning-Driven Discovery and Dynamics Simulation of Novel SQLE Inhibitors with ADMET Analysis and Molecular Modification | 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 Active Learning-Driven Discovery and Dynamics Simulation of Novel SQLE Inhibitors with ADMET Analysis and Molecular Modification YunYun Xu, Qiang Wang, GaoQiang Xu, YouJian Xu, Yiping Mou This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4163089/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 In this study, we leveraged a sophisticated active learning model to enhance virtual screening for SQLE inhibitors. The model's improved predictive accuracy identified compounds with significant advantages in binding affinity and thermodynamic stability. Detailed analyses, including molecular dynamics simulations and ADMET profiling, were conducted, particularly focusing on compounds CMNPD11566 and its derivative HCJ007. CMNPD11566 showed stable interactions with SQLE, but HCJ007 exhibited enhanced dynamic characteristics and interaction patterns. ADMET data comparison highlighted HCJ007's superior profile in terms of lower toxicity and better drug-likeness. Our findings suggest HCJ007 as a promising candidate for SQLE inhibition, with significant improvements over CMNPD11566 in various pharmacokinetic and safety parameters. The study underscores the efficacy of computational models in drug discovery and the importance of comprehensive preclinical evaluations. Biological sciences/Cancer/Cancer prevention Biological sciences/Cancer/Cancer therapy Marine Natural Products Screening of SQLE Inhibitors Active Learning Model ADMET Analysis Molecular Dynamics Simulations Molecular Modification Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 1. Introduction The relentless challenge posed by pancreatic cancer, with its high mortality rate and limited treatment options, underscores the urgent need for innovative therapeutic strategies 1 . The emerging role of cholesterol metabolism in cancer progression has opened new avenues for research 2 , particularly focusing on squalene epoxidase (SQLE). This enzyme, pivotal in cholesterol metabolism, is notably overexpressed in pancreatic cancer, presenting an intriguing therapeutic target 3 . SQLE's significant upregulation in pancreatic cancer tissues correlates with poor prognosis, as observed in recent studies 4 . This enzyme not only accelerates cell proliferation and cell cycle progression but also inhibits apoptosis, both in vitro and in vivo. SQLE's mechanistic action is twofold: it alleviates endoplasmic reticulum stress and activates the lipid raft-mediated Src/PI3K/Akt signaling pathway, thus propelling pancreatic cancer growth. The efficacy of SQLE inhibitors in reducing pancreatic cancer cell proliferation and impeding xenograft tumor growth further underscores SQLE's potential as a therapeutic target. Marine natural products (MNPs) serve as a crucial source for developing novel anticancer agents 5 . Due to their unique chemical structures and diverse biological activities, these compounds are viewed as a promising avenue for novel cancer therapies. For instance, iodinated carrageenan, isolated from red algae, has been employed in treating respiratory infections caused by rhinoviruses 6 . Trabectedin, extracted from the Caribbean sea squirt Ecteinascidia turbinate, is used for treating ovarian cancer and soft tissue sarcoma 7 . Brentuximab vedotin, derived from the marine mollusk Dolabella auricularia, is utilized in the treatment of Hodgkin lymphoma and large cell lymphoma 8 . Another example is Ziconotide, isolated from the venom of the marine snail Conus magnus, which is an effective analgesic, a thousand times more potent than morphine, and hence approved for pain management in HIV and cancer patients 9 . Overall, the exploration of marine natural products offers a rich resource and potential new pathways for developing novel anticancer drugs, especially in unraveling the complex and varied mechanisms of cancer diseases. With advancements in technology and deeper understanding of marine biodiversity, more unique marine natural products with therapeutic potential are expected to be discovered 10 . The integration of computational biology in drug discovery, particularly in screening marine natural products, marks a significant shift in identifying new therapeutic agents 11 . Utilizing advanced computational tools like active learning algorithms, molecular dynamics simulations, and diverse scoring systems, we can efficiently explore vast libraries of marine natural products 12 . This approach allows for the rapid and precise identification of potential SQLE inhibitors, predicting their effectiveness and safety profiles with greater accuracy. This study aims to leverage the chemical diversity of marine natural products to identify novel SQLE inhibitors for treating pancreatic cancer. By adopting an integrative computational strategy, we aim to methodically screen a marine natural products library, uncovering promising SQLE inhibitors. This endeavor is not only pivotal in advancing pancreatic cancer therapeutics but also showcases the vast potential of marine natural products in drug discovery, particularly when augmented by computational biology techniques. 2. Results 2.1 Application of active learning model in the screening of potential SQLE inhibitors Leveraging an active learning model, we have significantly enhanced the precision of virtual screening protocols for the identification of SQLE inhibitors, methodically detailed in Figure 1. The evolved iterations of the model, as visualized in Figure 1A-D, demonstrate an upward trajectory in predictive accuracy, with R² values advancing from 0.52 to a commendable 0.68. This increment evidences the model's augmented capacity to discern bona fide inhibitors amidst a complex chemical library.In Figure 1E, the spatial distribution of the shortlisted inhibitors within the SQLE binding pocket underscores the model's acuity in pinpointing molecules that exhibit structural compatibility at the active site. Figure 1F offers an incisive interaction fingerprint analysis, elucidating the nature of the contacts between the inhibitors and the critical amino acid residues of SQLE. This analysis is instrumental for the prospective structural refinement of these molecular entities. Extending this analysis, as shown in Table 1, we note that the selected compounds uniformly present a state penalty of zero, indicative of their unerring alignment with the target enzyme's active site sans promiscuous off-target interactions. Their docking scores oscillate between -10.036 to -11.644, decisively besting the benchmarks set by known inhibitors and signifying a potent binding affinity. The ligand strain energy metrics remain comfortably beneath the 10 kcal/mol ceiling, suggesting an innate propensity for these compounds to adopt a conformationally relaxed pose within the enzyme's pocket. Moreover, the MMGBSA dG Bind values, deeply ensconced in the negative spectrum from -60.47 to -73.52 kcal/mol, are profoundly indicative of thermodynamically stable complexes with the enzyme, particularly for compound 18775 which exhibits the most negative binding energy. This stark contrast with the MMGBSA values of established inhibitors, spanning from -49.18 to -60.66 kcal/mol, positions the screened compounds as potentially superior contenders, endowed with substantial binding free energy favoring stable and potent inhibition. In summation, the compounds emerging from this enhanced computational screening process, armed with formidable docking scores, minimal ligand strain energies, and profoundly negative MMGBSA dG Bind values, stand poised to redefine efficacy in SQLE inhibition. They represent not only a leap forward in virtual screening outcomes but also a promising stride towards the discovery of new therapeutic leads that could surpass the inhibitory capabilities of current pharmacological agents. 2.2 Analysis of compound interactions and scoring in BPMD simulations Binding Pose Metadynamics (BPMD), an advanced protocol harnessing automated and enhanced sampling based on metadynamics principles, effectively facilitates the dynamic exploration of ligand binding poses. This innovative approach has shown proficiency in distinguishing the accurate binding pose of ligands from less precise alternatives often generated in conventional docking studies. Utilizing BPMD, we analyzed the conformational stability of thirteen compounds over ten simulations, each lasting 10 nanoseconds. The collective variable root mean square deviation (CV RMSD) profiles focused on compounds maintaining stability within a 2 Å fluctuation threshold. The associated scoring metrics - PoseScore, PersScore, and CompScore - played a crucial role in refining the selection process. A lower PoseScore indicates enhanced stability, while a higher PersScore suggests the sustained presence of hydrogen bonds. CompScore, integrating these factors, identifies complexes with the highest stability. Within this framework, compounds CMNPD8663, CMNPD8904, CMNPD11566, and CMNPD18677 distinguished themselves with optimal stability profiles, as shown in Figure 2A. CMNPD8663, as illustrated in Figure 2B, primarily relies on hydrophobic interactions within the SQLE receptor's pocket. Despite the absence of specific hydrogen bonds, these interactions form the cornerstone of the compound's affinity. CMNPD8904, depicted in Figure 2C, exhibits a similar pattern of hydrophobic interactions, with potential Pi-Pi stacking with Tyr-195 contributing to enhanced binding specificity. CMNPD11566, shown in Figure 2D, displays a more intricate interaction profile, combining hydrophobic contacts and hydrogen bonds with Tyr-195 and Leu-416, indicating a refined interaction that augments both stability and specificity. Finally, CMNPD18677 (Figure 2E) presents a multifaceted binding mode, incorporating extensive hydrophobic interactions, a salt bridge with Tyr-195, and hydrogen bonds with Leu-416 and Thr-417, indicative of high affinity and selectivity. These observations provide a comprehensive understanding of the diverse interactions between ligands and the SQLE receptor, emphasizing the complex interplay of hydrophobic forces, hydrogen bonding, and salt bridge formation. Specifically, CMNPD11566 and CMNPD18677 exhibit advanced binding properties, suggesting their potential as effective SQLE inhibitors. 2.3 Evaluation of physicochemical properties and drug-likeness criteria for selected compounds ADMET analysis serves as a cornerstone in drug development, offering vital insights into how a compound interacts with the body. This encompasses absorption, distribution, metabolism, excretion, and toxicity—factors that collectively dictate a compound's success as a drug. Within the context of physicochemical properties, depicted in Figure 3A-D, the compounds CMNPD8663, CMNPD8904, CMNPD11566, and CMNPD18677 showcase distinct profiles. CMNPD8663 and CMNPD8904 display moderate physicochemical attributes conducive to good permeability and distribution. CMNPD11566 presents a particularly advantageous profile, with a lower molecular weight and topological polar surface area hinting at superior oral bioavailability. CMNPD18677, with its broader property spread, suggests a capacity for more complex interactions within biological systems. Transitioning to drug-likeness, Figure 3E summarizes the compliance of these compounds with established drug-likeness rules. While CMNPD8663 falls short on several parameters, CMNPD8904 is aligned with Lipinski's Rule, indicating its potential as an orally available drug. Notably, CMNPD11566 stands out, conforming to all the applied rules, including Lipinski, Pfizer, GSK, and the Golden Triangle, marking it as the most promising candidate in terms of drug-likeness. Focusing on CMNPD11566, which emerges as the front-runner in drug-likeness, its ADMET characteristics are explored in Supplementary Table 1. Although the Caco-2 permeability data suggest some absorption challenges, the MDCK permeability values offer a counterbalance, indicating the possibility of effective passive diffusion. The high plasma protein binding percentage points to a significant bound fraction in circulation, with only a small percentage available for active pharmacological effect. Metabolically, the compound engages with a variety of Cytochrome P450 enzymes, a factor that requires careful monitoring due to the potential for drug-drug interactions. The clearance and half-life data of CMNPD11566 imply a moderate rate of elimination and a brief systemic presence, respectively, influencing dosing intervals. Toxicity predictions, though varied, signal the need for careful consideration of the compound's safety profile, especially in areas of genotoxic carcinogenicity and skin sensitization. In conclusion, while the physicochemical properties of CMNPD11566 provide a favorable starting point, and its adherence to drug-likeness rules is exemplary, the comprehensive ADMET profile underscores the need for a nuanced approach to its development. Thorough preclinical studies will be essential to navigate its complex ADMET landscape and fully elucidate its therapeutic potential. 2.4 Molecular Dynamics Simulation and Biophysical Analysis of CMNPD11566 and Its Derivative HCJ007 in Interaction with SQLE Molecular dynamics simulations are essential in analyzing the structure-activity relationships of drug candidates, providing critical insights into their interactions with target proteins. In the 1000 ns molecular dynamics simulation analysis of CMNPD11566, as illustrated in Figures 4A-C, we identified several potential limitations. Figure 4A displays the RMSD trajectories of CMNPD11566, with fluctuations ranging between 1 to 3 Å. Although this indicates relatively stable interactions with the SQLE enzyme, there is still room for further optimization. Figure 4B shows that while CMNPD11566 has high-frequency interactions with residues like Tyr-195 and Thr-417, it also exhibits extremely low interaction frequencies with residues such as Met-196 and Ile-197. This suggests that the structure-activity relationship of CMNPD11566 with SQLE could potentially be compromised by its inherent instability. Further, Figure 4C reveals the reasons behind many low-frequency interaction sites of CMNPD11566; notably, its alkyl chain fails to form stable interactions with residues in the pocket. Given the mobility of CMNPD11566's alkyl chain, which could lead to binding stability issues, we deemed structural modification necessary. By carefully analyzing the interaction patterns of CMNPD11566 and incorporating structural insights from known target inhibitors Naftifine and Butenafine, we designed a new compound, HCJ007. This approach aimed to address the identified binding issues of CMNPD11566 while preserving its effective interactions with key residues. Subsequent simulations of HCJ007, as depicted in Figures 4D-F, showed significant improvements in its dynamic characteristics. Figure 4D reveals a more stable RMSD for HCJ007, indicating enhanced binding conformation stability. Figure 4E shows that HCJ007 has increased interaction frequencies with key residues, and significantly fewer low-frequency interaction sites. Figure 4F further illustrates the detailed interaction patterns of HCJ007, revealing stronger and more consistent binding interactions compared to CMNPD11566, suggesting that it may have more effective binding and therapeutic action. In orchestrating the development of HCJ007 as a potent therapeutic against SQLE, a series of intricate biophysical analyses were pivotal, each unraveling a layer of interaction dynamics and molecular stability. Throughout its modification process, HCJ007 consistently adhered to the four key drug-likeness rules, and notably, it displayed no PAINS (Pan-Assay Interference Compounds) alerts, as detailed in Supplementary Table 2. These insights are instrumental in sculpting HCJ007's design, focusing on enhancing its specificity and efficacy. Initially, Figure 5A reveals the evolving affinity between HCJ007 and SQLE through the lens of MMGBSA binding free energy. The trend of deepening negative values, averaging at -124.48 kcal/mol and bottoming out at -151.82 kcal/mol, illustrates an intensifying molecular interaction. This trend is not a mere data trajectory but underscores HCJ007's potential as a high-affinity inhibitor, robustly engaging with SQLE.Building upon this overarching energy landscape, Figure 5B delves into the molecular intricacies, mapping the contribution of individual residues. The marked negative contributions of residues like Asn-193 and Asn-382 transcend being mere numerical values; they are pivotal in stabilizing the HCJ007-SQLE complex, guiding the precision in drug design and optimization. Transitioning to the atomic scale, Figure 5C offers insights into HCJ007's internal dynamics. Here, the RMSF increment starting from atom 16 suggests a deliberate retention of flexibility within the alkyl chain, essential for maintaining an adaptable binding conformation.This narrative of internal mobility and conformational adaptability is further complemented by Figure 5D's torsion angle analysis. The minimal fluctuations in torsion angles lend credence to the structural integrity of the alkyl chain, ensuring its functional alignment during the binding interaction. Segueing into the realm of molecular stability and interface dynamics, Figures 5E and 5F together illustrate a cohesive story. Figure 5E showcases the consistent pattern of intra-molecular hydrogen bonding, a hallmark of molecular stability and functional integrity. The steadiness punctuated by transient spikes mirrors the compound's ability to dynamically respond to structural shifts within the complex, indicating a resilient yet adaptable binding profile. Complementing this, Figure 5F shifts the focus to the broader interface properties like SASA, MolSA, and PSA. Their concurrent trends, subtly modulating the hydrophobic-hydrophilic balance, narrate an intricate tale of the molecular interface. These forces are not merely coexisting; they are intricately intertwined, dictating the binding dynamics and stability of the molecular complex. In totality, these analytical vignettes paint a composite picture of HCJ007 as a meticulously tuned compound for efficacy. Each atomic and molecular interaction within HCJ007 is choreographed to ensure synergistic engagement with SQLE, setting the stage for HCJ007's evolution as a promising therapeutic agent. 3. Discussion Leveraging the diverse chemical landscape of marine-derived compounds, this study has identified promising squalene epoxidase (SQLE) inhibitors, advancing the search for effective pancreatic cancer treatments. The application of an active learning model has refined virtual screening processes, pinpointing compounds with high binding affinity and requisite stability within the SQLE active site. This innovative approach has enabled the discovery of compounds that not only exhibit high theoretical efficacy but also hold potential for significant advancements in SQLE inhibition. Building on this foundation, our comprehensive analysis using multiple scoring criteria for the thirteen compounds has shown significant advantages in various aspects. Notably, the docking scores of all selected compounds exceeded -10, with CMNPD18677 achieving -11.644, indicating exceptionally strong binding affinity with SQLE 13 . This is further corroborated by the state penalty of zero for all compounds, emphasizing their high specificity in binding to SQLE and reducing the likelihood of non-specific interactions 14 . Additionally, ligand strain energy is a crucial metric for assessing how a ligand's energy changes when binding to a target protein 15 . Ideally, lower ligand strain energy means that the ligand can more naturally adapt to the protein's binding pocket, reducing the energy loss caused by conformational changes and increasing binding affinity. In our screening results, most selected compounds exhibited ligand strain energy values below 10 kcal/mol, significantly lower than known inhibitors like Butenafine (17.45 kcal/mol), Naftifine (9.666 kcal/mol), and Terbinafine (8.458 kcal/mol). For instance, compound 18775 had a strain energy of only 4.051 kcal/mol, much lower than Butenafine. This suggests that these compounds may bind more easily to SQLE, as they require less conformational change. These findings highlight the effectiveness of using an active learning model to select marine natural products as SQLE inhibitors. Compounds with low strain energy are theoretically more suitable as drug candidates, as they demand less energy to bind to their target proteins, potentially enhancing their efficacy and stability within the body. Most impressively, the MMGBSA dG Bind scores reveal that all selected compounds have significantly lower scores compared to known inhibitors. Particularly remarkable is CMNPD18775, which scored -73.52 kcal/mol, far surpassing the scores of Butenafine, Naftifine, and Terbinafine. This indicates that the complexes formed with SQLE are likely to be more thermodynamically stable and potentially more effective as inhibitors 16 . This finding underscores the effectiveness of the active learning model in efficiently identifying potent and stable SQLE inhibitors from a vast repository of marine-derived compounds. Transitioning to the dynamic characteristics, CMNPD11566 exhibited limitations in its interaction dynamics with SQLE, as seen in the molecular dynamics simulations. Its RMSD trajectories showed fluctuations within the range of 1-3 Å, suggesting areas for optimization despite relatively stable interactions 17 . The interaction frequency analysis revealed high-frequency interactions with residues like Tyr-195 and Thr-417 but lower interaction frequencies with others, indicating potential instability and suboptimal interaction with the enzyme's binding pocket. Tyr-195 was identified as a crucial residue in the SQLE enzyme. This residue plays a significant role in inhibitor binding, as evidenced by the fact that both NB-598 18 and Cmpd-4" establish a hydrogen bond with Tyr-195, which is the only specific and directional interaction these compounds have with SQLE 19 . This interaction is consistent across all known SQLE inhibitors and explains the required presence of the tertiary amine motif in these inhibitors. The interaction with conserved Tyr-195 is critical for the binding and efficacy of SQLE inhibitors. In contrast, HCJ007 displayed markedly improved dynamic characteristics. Its more stable RMSD, enhanced binding conformation stability, and increased interaction frequencies with key residues, as seen in subsequent simulations, highlight its potential for more effective binding and therapeutic action. This comparison underscores the effectiveness of the structural modifications made to HCJ007, resulting in improved interaction dynamics essential for successful SQLE inhibition. Finally, the ADMET data comparison reveals that HCJ007 offers significant improvements over CMNPD11566 in various parameters, such as lower BBB penetration and PPB, suggesting better central nervous system accessibility and higher active drug circulation 20 . Despite a higher clearance rate, HCJ007's comparable half-life to CMNPD11566 suggests similar durations in the body. Additionally, HCJ007 shows significantly reduced toxicity in AMES test results and skin sensitization, lower risks of hERG cardiac toxicity and DILI, and potentially better drug-like properties as indicated by its higher QED score. Based on these findings, the study propels us into new realms of research and clinical development, particularly focusing on the potential of HCJ007 as a SQLE inhibitor. With its superior dynamic characteristics and significant ADMET advantages over CMNPD11566, particularly in toxicity, safety, and drug-likeness, HCJ007 stands as a strong candidate for further exploration. This research lays a foundation for innovative approaches in drug discovery, combining computational modeling with empirical data to refine drug efficacy and safety profiles. The next steps involve validating these results through in vitro and in vivo studies, aiming towards eventual clinical trials. Such progression towards translational research emphasizes the crucial role of comprehensive preclinical evaluations, ensuring that compounds like HCJ007 are not only effective in laboratory settings but also well-suited for real-world therapeutic applications. 4. Materials and Methods 4.1 Protein Preparation In our quest to discover inhibitors targeting SQLE, we adopted structure-based computational techniques, focusing on the FGFR3 crystal structure (PDB ID: 6C6N) complexed with Cmpd-4" 19 . This particular structure was chosen to serve as the receptor model. The preparation of the receptor-ligand complex was carried out using Schrödinger’s Protein Preparation Wizard, which involved a series of meticulous steps. These included the addition of missing hydrogen atoms, adjustment of metal ion states, bond order determination in HET groups, assessment and optimization of ligand protonation states along with their energy implications, tuning protonation states of histidine residues, correcting any misplaced heavy atoms, refining the hydrogen bonding network within the protein, and executing a restrained minimization to ensure structural integrity.The identified binding site within the 3D structure of the receptor, where Cmpd-4" interacts, was designated as the focal point for screening potential ligands. Consequently, a grid corresponding to this target site was generated to facilitate the screening process. 4.2 Active learning based virtual screening Active Learning Glide will generate a receptor grid from a prepared protein, prepare the Comprehensive Marine Natural Products Database (https://cmnpd.org/), and dock a subset of these ligands using Glide SP 21 . Active Learning workflows train a machine learning (ML) model on physics-based data, such as FEP+ 22 predicted affinities or Glide docking scores, iteratively sampled from a full library using Schrödinger's deep-learning powered QSAR platform, DeepAutoQSAR (https://www.schrodinger.com/science-articles/benchmark-study-deepautoqsar-chemprop-and-deeppurpose-admet-subset-therapeutic-data). 3 iterative training rounds were set. After all the ligands have been screened using the last model, a selection of the top ligands will then be docked using Glide SP. The results of the docking were then quantified based on the consensus of docking scores and Prime MM-GBSA energy 23,24 . 4.3 Binding pose metadynamics To verify the stability of binding poses of the selected ligands from molecular dynamics (MD) simulations in both binding sites, a sequence of metadynamics MD simulations was executed, each lasting for 10 ns, on various docked poses and MD-established stable protein-ligand complexes 25–27 . The chosen collective variable (CV) was the root mean square deviation (RMSD) of the ligand's heavy atoms from their initial positions, computed post alignment of the binding sites to mitigate any drift. The parameters for the hill's height and width in the metadynamics simulations were set at 0.05 kcal/mol and 0.02 Å, respectively. The system underwent solvation in a box with a 10 Å buffer, followed by a series of minimization steps, gradually elevating the system's temperature to 300 K and alleviating any initial structural stresses or contacts. The evaluation of stability was based on monitoring the RMSD fluctuations of the ligand throughout the simulation (termed as PoseScore) and quantifying the average duration of critical contacts between the ligand and protein residues (referred to as PersScore). 4.4 ADMET Screening and Drug-Likeness Predictions To ensure favorable ADMET (absorption, distribution, metabolism, excretion, and toxicity) properties and non-toxicity of potential drug candidates, the study employed ADMETLAB 2.0 (https://admetmesh.scbdd.com/) 28 for prediction. Adherence to medicinal chemistry guidelines, such as the Lipinski 29 , Pfizer 30 , GSK 31 , and Golden Triangle rules 32 , was pivotal in identifying compounds with optimal ADMET attributes. 4.5 Molecular dynamic simulation The Desmond software from Schrödinger LLC was employed for conducting molecular dynamics (MD) simulations over a period of 1000 ns. The TIP3P water model, representing a three-point intermolecular interaction potential, was chosen for these simulations. The simulations were set up within an orthorhombic box, maintaining a constant temperature of 300 K and pressure of 1 atm, while utilizing the OPLS 2005 force field 33 . To achieve neutrality in the models and mimic physiological conditions, counter ions were added, and a 0.15 M sodium chloride solution was used to simulate ligand-binding status in physiological environment 34 . Prior to the commencement of the simulations, the models underwent a relaxation phase, and during the simulations, data was recorded and stored every 100 ps for later analysis. 5. Conclusions This study marks a significant advancement in identifying potent squalene epoxidase (SQLE) inhibitors for pancreatic cancer treatment. Utilizing an active learning model, we identified compounds, notably HCJ007, with superior binding affinity, stability, and ADMET properties compared to known inhibitors. HCJ007's enhanced dynamic characteristics and reduced toxicity profiles position it as a promising therapeutic candidate. These findings pave the way for further validation through in vitro and in vivo studies, potentially leading to clinical trials and offering new avenues in cancer treatment. Declarations Data availability: All data generated or analysed during this study are included in this published article [and its supplementary information files]. 8.Author Contributions: Conceptualization, YiPing Mou; Data curation, Yunyun Xu and YouJian Xu; Funding acquisition, YiPing Mou; Project administration, YiPing Mou; Resources, Qiang Wang; Validation, Yunyun Xu; Writing – original draft, Yunyun Xu; Writing – review & editing, YiPing Mou. All authors read and approved the final manuscript. 9. Funding: This research was funded by Scientific Research Fund of Zhejiang Provincial Education Department, grant number Y202249377. Conflicts of Interest: The authors declare no conflict of interest. References Halbrook, C. J., Lyssiotis, C. A., Pasca di Magliano, M. & Maitra, A. Pancreatic cancer: Advances and challenges. Cell 186, 1729–1754 (2023). Xiao, M. et al. Functional significance of cholesterol metabolism in cancer: from threat to treatment. Exp Mol Med 55, 1982–1995 (2023). You, W. et al. SQLE, A Key Enzyme in Cholesterol Metabolism, Correlates With Tumor Immune Infiltration and Immunotherapy Outcome of Pancreatic Adenocarcinoma. Front Immunol 13, 864244 (2022). Xu, R. et al. SQLE promotes pancreatic cancer growth by attenuating ER stress and activating lipid rafts-regulated Src/PI3K/Akt signaling pathway. Cell Death Dis 14, 497 (2023). Khalifa, S. A. M. et al. Marine Natural Products: A Source of Novel Anticancer Drugs. Mar Drugs 17, 491 (2019). Frediansyah, A. The antiviral activity of iota-, kappa-, and lambda-carrageenan against COVID-19: A critical review. Clin Epidemiol Glob Health 12, 100826 (2021). Carter, N. J. & Keam, S. J. Trabectedin: a review of its use in soft tissue sarcoma and ovarian cancer. Drugs 70, 355–376 (2010). Donato, E. M., Fernández-Zarzoso, M., Hueso, J. A. & de la Rubia, J. Brentuximab vedotin in Hodgkin lymphoma and anaplastic large-cell lymphoma: an evidence-based review. Onco Targets Ther 11, 4583–4590 (2018). Banik, R. K. & Engle, M. P. Ziconotide for Management of Cancer Pain Refractory to Pharmacotherapy: An Update. Pain Med 21, 3253–3259 (2020). McConnell, O. J., Longley, R. E. & Koehn, F. E. The discovery of marine natural products with therapeutic potential. Biotechnology 26, 109–174 (1994). Abduljalil, J. M., Elfiky, A. A., Sayed, E.-S. T. A. & AlKhazindar, M. M. Computational identification of drug-like marine natural products as potential RNA polymerase inhibitors against Nipah virus. Comput Biol Chem 104, 107850 (2023). Tran, D. P. et al. Using molecular dynamics simulations to prioritize and understand AI-generated cell penetrating peptides. Sci Rep 11, 10630 (2021). Meli, R., Morris, G. M. & Biggin, P. C. Scoring Functions for Protein-Ligand Binding Affinity Prediction using Structure-Based Deep Learning: A Review. Front Bioinform 2, 885983 (2022). Friesner, R. A. et al. Extra precision glide: docking and scoring incorporating a model of hydrophobic enclosure for protein-ligand complexes. J Med Chem 49, 6177–6196 (2006). Tong, J. & Zhao, S. Large-Scale Analysis of Bioactive Ligand Conformational Strain Energy by Ab Initio Calculation. J Chem Inf Model 61, 1180–1192 (2021). Pattar, S., Adhoni, S. A., Kamanavalli, C. M. & Kumbar, S. S. In silico molecular docking studies and MM/GBSA analysis of coumarin-carbonodithioate hybrid derivatives divulge the anticancer potential against breast cancer. Beni-Suef University Journal of Basic and Applied Sciences 9, (2020). Kumar, M. et al. Definition of fatty acid transport protein-2 (FATP2) structure facilitates identification of small molecule inhibitors for the treatment of diabetic complications. Int J Biol Macromol 244, 125328 (2023). Horie, M. et al. NB-598: a potent competitive inhibitor of squalene epoxidase. J Biol Chem 265, 18075–18078 (1990). Padyana, A. K. et al. Structure and inhibition mechanism of the catalytic domain of human squalene epoxidase. Nat Commun 10, 97 (2019). Tong, X. et al. Blood-brain barrier penetration prediction enhanced by uncertainty estimation. J Cheminform 14, 44 (2022). Friesner, R. A. et al. Glide: a new approach for rapid, accurate docking and scoring. 1. Method and assessment of docking accuracy. J Med Chem 47, 1739–1749 (2004). Wang, L. et al. Accurate and reliable prediction of relative ligand binding potency in prospective drug discovery by way of a modern free-energy calculation protocol and force field. J Am Chem Soc 137, 2695–2703 (2015). Mittal, L., Kumari, A., Suri, C., Bhattacharya, S. & Asthana, S. Insights into structural dynamics of allosteric binding sites in HCV RNA-dependent RNA polymerase. J Biomol Struct Dyn 38, 1612–1625 (2020). Mittal, L., Kumari, A., Srivastava, M., Singh, M. & Asthana, S. Identification of potential molecules against COVID-19 main protease through structure-guided virtual screening approach. J Biomol Struct Dyn 39, 3662–3680 (2021). Clark, A. J. et al. Prediction of Protein-Ligand Binding Poses via a Combination of Induced Fit Docking and Metadynamics Simulations. J Chem Theory Comput 12, 2990–2998 (2016). Fusani, L., Palmer, D. S., Somers, D. O. & Wall, I. D. Exploring Ligand Stability in Protein Crystal Structures Using Binding Pose Metadynamics. J Chem Inf Model 60, 1528–1539 (2020). Purushotham, N. et al. Design and synthesis of amino acid derivatives of substituted benzimidazoles and pyrazoles as Sirt1 inhibitors. RSC Adv 12, 3809–3827 (2022). Xiong, G. et al. ADMETlab 2.0: an integrated online platform for accurate and comprehensive predictions of ADMET properties. Nucleic Acids Res 49, W5–W14 (2021). Lipinski, C. A., Lombardo, F., Dominy, B. W. & Feeney, P. J. Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. Adv Drug Deliv Rev 46, 3–26 (2001). Hughes, J. D. et al. Physiochemical drug properties associated with in vivo toxicological outcomes. Bioorg Med Chem Lett 18, 4872–4875 (2008). Gleeson, M. P. Generation of a set of simple, interpretable ADMET rules of thumb. J Med Chem 51, 817–834 (2008). Johnson, T. W., Dress, K. R. & Edwards, M. Using the Golden Triangle to optimize clearance and oral absorption. Bioorg Med Chem Lett 19, 5560–5564 (2009). Shivakumar, D. et al. Prediction of Absolute Solvation Free Energies using Molecular Dynamics Free Energy Perturbation and the OPLS Force Field. J Chem Theory Comput 6, 1509–1519 (2010). Rasheed, M. A. et al. Identification of Lead Compounds against Scm (fms10) in Enterococcus faecium Using Computer Aided Drug Designing. Life (Basel) 11, 77 (2021). Table Table 1 is available in the Supplementary Files section. Additional Declarations No competing interests reported. Supplementary Files Tab1.csv Suptab1.csv Suptab2.csv 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-4163089","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":291546311,"identity":"e6d01dd7-bae9-4bf2-b78d-a72600329914","order_by":0,"name":"YunYun Xu","email":"","orcid":"","institution":"Zhejiang Provincial People's Hospital (Affiliated People's Hospital), Hangzhou Medical College","correspondingAuthor":false,"prefix":"","firstName":"YunYun","middleName":"","lastName":"Xu","suffix":""},{"id":291546312,"identity":"ba2a7b62-a3ae-41b2-9037-70cea53f69c9","order_by":1,"name":"Qiang Wang","email":"","orcid":"","institution":"Tiantai People's Hospital","correspondingAuthor":false,"prefix":"","firstName":"Qiang","middleName":"","lastName":"Wang","suffix":""},{"id":291546314,"identity":"3d50f223-8370-4aa1-9a0d-09f1a79c2d9d","order_by":2,"name":"GaoQiang Xu","email":"","orcid":"","institution":"Tiantai People's Hospital","correspondingAuthor":false,"prefix":"","firstName":"GaoQiang","middleName":"","lastName":"Xu","suffix":""},{"id":291546317,"identity":"ca21cc8b-cb86-4c1a-a000-99af3037339a","order_by":3,"name":"YouJian Xu","email":"","orcid":"","institution":"Tiantai People's Hospital","correspondingAuthor":false,"prefix":"","firstName":"YouJian","middleName":"","lastName":"Xu","suffix":""},{"id":291546318,"identity":"30f93141-7bc7-4784-876c-9717637b905c","order_by":4,"name":"Yiping Mou","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAxElEQVRIiWNgGAWjYDACZgaGA2AGe2Pjgw+kaeE53Gw4gzTrJNLbpDmIUWhwnPfg4YJf2+QMbj5skGZgsJPTbSCk5TBfwuGZfbeNDW4nNhgXMCQbmx0gqIXH4DBvz+3EDUAtyTMYDiRuI17LzYMNh3mI1sLzA6jlBmNjM1FaJEF+4W24bSx5JrGZcYYBEX7hO3/28GeeP7fl+I4ff/7jQ4WdHEEtCgd4GBgY2+DuJKAcBOQbgFoY/hChchSMglEwCkYuAADICEwv5OuoNQAAAABJRU5ErkJggg==","orcid":"","institution":"Zhejiang Provincial People's Hospital (Affiliated People's Hospital), Hangzhou Medical College","correspondingAuthor":true,"prefix":"","firstName":"Yiping","middleName":"","lastName":"Mou","suffix":""}],"badges":[],"createdAt":"2024-03-25 12:00:41","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4163089/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4163089/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":54748453,"identity":"666e278f-5205-4f3c-971b-917e76f0966a","added_by":"auto","created_at":"2024-04-16 07:56:52","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":114556,"visible":true,"origin":"","legend":"\u003cp\u003eApplication of active learning model in the screening of potentional SQLE inhibitors: (a) Training results of the active learning model at different iterative stages. (b) Enrichment effect of SQLE target compounds after active learning combined with multidimensional scoring. (c) Interaction fingerprints of 13 selected compounds with SQLE.\u003c/p\u003e","description":"","filename":"1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4163089/v1/bcc8074f76bc706fff6b2721.jpg"},{"id":54748454,"identity":"e48242b9-17db-4347-ae4f-ea5bd5054816","added_by":"auto","created_at":"2024-04-16 07:56:52","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":118988,"visible":true,"origin":"","legend":"\u003cp\u003eAnalysis of compound interactions and scoring in BPMD simulations: (a) Left: Dynamic changes in CV RMSD over time for 13 compounds during 10 ns BPMD simulations repeated 10 times. Right: Three different scoring scenarios for the 13 compounds. (a-e) Four compounds selected through BPMD screening: CMNPD8663, CMNPD8904, CMNPD11566, and CMNPD18677, respectively.\u003c/p\u003e","description":"","filename":"2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4163089/v1/9bc72f5052a5ad72ac52b117.jpg"},{"id":54748456,"identity":"472dc6db-a7aa-4605-b9b8-9999795f4d52","added_by":"auto","created_at":"2024-04-16 07:56:52","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":102471,"visible":true,"origin":"","legend":"\u003cp\u003eEvaluation of physicochemical properties and drug-likeness criteria for selected compounds: (a-d) Radar charts of physicochemical properties for four compounds: CMNPD8663, CMNPD8904, CMNPD11566, and CMNPD18677, respectively. (e) Compliance with Lipinski's rule of five and the number of PAINS (Pan-Assay Interference Compounds) alerts for the four compounds.\u003c/p\u003e","description":"","filename":"3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4163089/v1/3966fec7f88caaff878f7a6f.jpg"},{"id":54748457,"identity":"8001b48a-affb-4019-9ad6-7a7e20785e62","added_by":"auto","created_at":"2024-04-16 07:56:52","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":137594,"visible":true,"origin":"","legend":"\u003cp\u003eMolecular dynamics simulation analysis of CMNPD11566 and its derivative HCJ007: (a-c) Simulations for CMNPD11566: (a) Three types of RMSD for CMNPD11566. (b) Interaction statistics of CMNPD11566. (c) 2D interaction diagram of CMNPD11566 with residues having interaction frequency over 30%. (d-f) Simulations for the derivative HCJ007: (d) Three types of RMSD for HCJ007. (e) Interaction statistics of HCJ007. (f) 2D interaction diagram of HCJ007 with residues having interaction frequency over 30%.\u003c/p\u003e","description":"","filename":"4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4163089/v1/1de2c4b3ba3648c5bb7e52f8.jpg"},{"id":54748455,"identity":"a97af4bd-d4b4-451f-81de-5523051511b8","added_by":"auto","created_at":"2024-04-16 07:56:52","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":126442,"visible":true,"origin":"","legend":"\u003cp\u003eComprehensive biophysical analysis of HCJ007 in interaction with SQLE:(a) MMGBSA binding free energy between HCJ007 and SQLE. (b) Contribution of individual residues to the HCJ007-SQLE complex stability. (c) RMSF analysis of HCJ007 (d) Torsion angle analysis of HCJ007. (e) Interface properties analysis of HCJ007, including SASA, MolSA, and PSA. (f) Intra-molecular hydrogen bonding patterns in HCJ007.\u003c/p\u003e","description":"","filename":"5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-4163089/v1/142fc68694e1179a107dcb82.jpg"},{"id":72889206,"identity":"43b51bf8-f967-4ebe-85ee-fe01699a0df1","added_by":"auto","created_at":"2025-01-03 10:32:14","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":905482,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4163089/v1/7d765660-8b34-48c5-9365-3542208a1972.pdf"},{"id":54748957,"identity":"8582b781-38dd-4fb1-b6f9-bd5a359af449","added_by":"auto","created_at":"2024-04-16 08:04:52","extension":"csv","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":446,"visible":true,"origin":"","legend":"","description":"","filename":"Tab1.csv","url":"https://assets-eu.researchsquare.com/files/rs-4163089/v1/c3e1eeb63fd9bcd02dd52c03.csv"},{"id":54748459,"identity":"77a3eed0-7735-4dcf-8cfe-0f6f4175e831","added_by":"auto","created_at":"2024-04-16 07:56:52","extension":"csv","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":1308,"visible":true,"origin":"","legend":"","description":"","filename":"Suptab1.csv","url":"https://assets-eu.researchsquare.com/files/rs-4163089/v1/1059756c97d0bb72c973eb08.csv"},{"id":54749358,"identity":"eb23877f-9e19-4e48-9b3c-c66ac3f0b699","added_by":"auto","created_at":"2024-04-16 08:12:52","extension":"csv","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":1301,"visible":true,"origin":"","legend":"","description":"","filename":"Suptab2.csv","url":"https://assets-eu.researchsquare.com/files/rs-4163089/v1/5a70815ad1d405685c96ffb2.csv"}],"financialInterests":"No competing interests reported.","formattedTitle":"Active Learning-Driven Discovery and Dynamics Simulation of Novel SQLE Inhibitors with ADMET Analysis and Molecular Modification","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eThe relentless challenge posed by pancreatic cancer, with its high mortality rate and limited treatment options, underscores the urgent need for innovative therapeutic strategies\u003csup\u003e1\u003c/sup\u003e. The emerging role of cholesterol metabolism in cancer progression has opened new avenues for research\u003csup\u003e2\u003c/sup\u003e, particularly focusing on squalene epoxidase (SQLE). This enzyme, pivotal in cholesterol metabolism, is notably overexpressed in pancreatic cancer, presenting an intriguing therapeutic target\u003csup\u003e3\u003c/sup\u003e. SQLE\u0026apos;s significant upregulation in pancreatic cancer tissues correlates with poor prognosis, as observed in recent studies\u003csup\u003e4\u003c/sup\u003e. This enzyme not only accelerates cell proliferation and cell cycle progression but also inhibits apoptosis, both in vitro and in vivo. SQLE\u0026apos;s mechanistic action is twofold: it alleviates endoplasmic reticulum stress and activates the lipid raft-mediated Src/PI3K/Akt signaling pathway, thus propelling pancreatic cancer growth. The efficacy of SQLE inhibitors in reducing pancreatic cancer cell proliferation and impeding xenograft tumor growth further underscores SQLE\u0026apos;s potential as a therapeutic target.\u003c/p\u003e\n\u003cp\u003eMarine natural products (MNPs) serve as a crucial source for developing novel anticancer agents\u003csup\u003e5\u003c/sup\u003e. Due to their unique chemical structures and diverse biological activities, these compounds are viewed as a promising avenue for novel cancer therapies. For instance, iodinated carrageenan, isolated from red algae, has been employed in treating respiratory infections caused by rhinoviruses\u003csup\u003e6\u003c/sup\u003e. Trabectedin, extracted from the Caribbean sea squirt Ecteinascidia turbinate, is used for treating ovarian cancer and soft tissue sarcoma\u003csup\u003e7\u003c/sup\u003e. Brentuximab vedotin, derived from the marine mollusk Dolabella auricularia, is utilized in the treatment of Hodgkin lymphoma and large cell lymphoma\u003csup\u003e8\u003c/sup\u003e. Another example is Ziconotide, isolated from the venom of the marine snail Conus magnus, which is an effective analgesic, a thousand times more potent than morphine, and hence approved for pain management in HIV and cancer patients\u003csup\u003e9\u003c/sup\u003e. Overall, the exploration of marine natural products offers a rich resource and potential new pathways for developing novel anticancer drugs, especially in unraveling the complex and varied mechanisms of cancer diseases.\u003c/p\u003e\n\u003cp\u003eWith advancements in technology and deeper understanding of marine biodiversity, more unique marine natural products with therapeutic potential are expected to be discovered\u003csup\u003e10\u003c/sup\u003e. The integration of computational biology in drug discovery, particularly in screening marine natural products, marks a significant shift in identifying new therapeutic agents\u003csup\u003e11\u003c/sup\u003e. Utilizing advanced computational tools like active learning algorithms, molecular dynamics simulations, and diverse scoring systems, we can efficiently explore vast libraries of marine natural products\u003csup\u003e12\u003c/sup\u003e. This approach allows for the rapid and precise identification of potential SQLE inhibitors, predicting their effectiveness and safety profiles with greater accuracy.\u003c/p\u003e\n\u003cp\u003eThis study aims to leverage the chemical diversity of marine natural products to identify novel SQLE inhibitors for treating pancreatic cancer. By adopting an integrative computational strategy, we aim to methodically screen a marine natural products library, uncovering promising SQLE inhibitors. This endeavor is not only pivotal in advancing pancreatic cancer therapeutics but also showcases the vast potential of marine natural products in drug discovery, particularly when augmented by computational biology techniques.\u003c/p\u003e"},{"header":"2. Results","content":"\u003cp\u003e2.1 Application of active learning model in the screening of potential SQLE inhibitors\u003c/p\u003e\n\u003cp\u003eLeveraging an active learning model, we have significantly enhanced the precision of virtual screening protocols for the identification of SQLE inhibitors, methodically detailed in Figure 1. The evolved iterations of the model, as visualized in Figure 1A-D, demonstrate an upward trajectory in predictive accuracy, with R\u0026sup2; values advancing from 0.52 to a commendable 0.68. This increment evidences the model\u0026apos;s augmented capacity to discern bona fide inhibitors amidst a complex chemical library.In Figure 1E, the spatial distribution of the shortlisted inhibitors within the SQLE binding pocket underscores the model\u0026apos;s acuity in pinpointing molecules that exhibit structural compatibility at the active site. Figure 1F offers an incisive interaction fingerprint analysis, elucidating the nature of the contacts between the inhibitors and the critical amino acid residues of SQLE. This analysis is instrumental for the prospective structural refinement of these molecular entities.\u003c/p\u003e\n\u003cp\u003eExtending this analysis, as shown in Table 1, we note that the selected compounds uniformly present a state penalty of zero, indicative of their unerring alignment with the target enzyme\u0026apos;s active site sans promiscuous off-target interactions. Their docking scores oscillate between -10.036 to -11.644, decisively besting the benchmarks set by known inhibitors and signifying a potent binding affinity. The ligand strain energy metrics remain comfortably beneath the 10 kcal/mol ceiling, suggesting an innate propensity for these compounds to adopt a conformationally relaxed pose within the enzyme\u0026apos;s pocket. Moreover, the MMGBSA dG Bind values, deeply ensconced in the negative spectrum from -60.47 to -73.52 kcal/mol, are profoundly indicative of thermodynamically stable complexes with the enzyme, particularly for compound 18775 which exhibits the most negative binding energy. This stark contrast with the MMGBSA values of established inhibitors, spanning from -49.18 to -60.66 kcal/mol, positions the screened compounds as potentially superior contenders, endowed with substantial binding free energy favoring stable and potent inhibition.\u003c/p\u003e\n\u003cp\u003eIn summation, the compounds emerging from this enhanced computational screening process, armed with formidable docking scores, minimal ligand strain energies, and profoundly negative MMGBSA dG Bind values, stand poised to redefine efficacy in SQLE inhibition. They represent not only a leap forward in virtual screening outcomes but also a promising stride towards the discovery of new therapeutic leads that could surpass the inhibitory capabilities of current pharmacological agents.\u003c/p\u003e\n\u003cp\u003e2.2 Analysis of compound interactions and scoring in BPMD simulations\u003c/p\u003e\n\u003cp\u003eBinding Pose Metadynamics (BPMD), an advanced protocol harnessing automated and enhanced sampling based on metadynamics principles, effectively facilitates the dynamic exploration of ligand binding poses. This innovative approach has shown proficiency in distinguishing the accurate binding pose of ligands from less precise alternatives often generated in conventional docking studies. Utilizing BPMD, we analyzed the conformational stability of thirteen compounds over ten simulations, each lasting 10 nanoseconds. The collective variable root mean square deviation (CV RMSD) profiles focused on compounds maintaining stability within a 2 \u0026Aring; fluctuation threshold. The associated scoring metrics - PoseScore, PersScore, and CompScore - played a crucial role in refining the selection process. A lower PoseScore indicates enhanced stability, while a higher PersScore suggests the sustained presence of hydrogen bonds. CompScore, integrating these factors, identifies complexes with the highest stability.\u003c/p\u003e\n\u003cp\u003eWithin this framework, compounds CMNPD8663, CMNPD8904, CMNPD11566, and CMNPD18677 distinguished themselves with optimal stability profiles, as shown in Figure 2A. CMNPD8663, as illustrated in Figure 2B, primarily relies on hydrophobic interactions within the SQLE receptor\u0026apos;s pocket. Despite the absence of specific hydrogen bonds, these interactions form the cornerstone of the compound\u0026apos;s affinity. CMNPD8904, depicted in Figure 2C, exhibits a similar pattern of hydrophobic interactions, with potential Pi-Pi stacking with Tyr-195 contributing to enhanced binding specificity. CMNPD11566, shown in Figure 2D, displays a more intricate interaction profile, combining hydrophobic contacts and hydrogen bonds with Tyr-195 and Leu-416, indicating a refined interaction that augments both stability and specificity. Finally, CMNPD18677 (Figure 2E) presents a multifaceted binding mode, incorporating extensive hydrophobic interactions, a salt bridge with Tyr-195, and hydrogen bonds with Leu-416 and Thr-417, indicative of high affinity and selectivity.\u003c/p\u003e\n\u003cp\u003eThese observations provide a comprehensive understanding of the diverse interactions between ligands and the SQLE receptor, emphasizing the complex interplay of hydrophobic forces, hydrogen bonding, and salt bridge formation. Specifically, CMNPD11566 and CMNPD18677 exhibit advanced binding properties, suggesting their potential as effective SQLE inhibitors.\u003c/p\u003e\n\u003cp\u003e2.3 Evaluation of physicochemical properties and drug-likeness criteria for selected compounds\u003c/p\u003e\n\u003cp\u003eADMET analysis serves as a cornerstone in drug development, offering vital insights into how a compound interacts with the body. This encompasses absorption, distribution, metabolism, excretion, and toxicity\u0026mdash;factors that collectively dictate a compound\u0026apos;s success as a drug.\u003c/p\u003e\n\u003cp\u003eWithin the context of physicochemical properties, depicted in Figure 3A-D, the compounds CMNPD8663, CMNPD8904, CMNPD11566, and CMNPD18677 showcase distinct profiles. CMNPD8663 and CMNPD8904 display moderate physicochemical attributes conducive to good permeability and distribution. CMNPD11566 presents a particularly advantageous profile, with a lower molecular weight and topological polar surface area hinting at superior oral bioavailability. CMNPD18677, with its broader property spread, suggests a capacity for more complex interactions within biological systems.\u003c/p\u003e\n\u003cp\u003eTransitioning to drug-likeness, Figure 3E summarizes the compliance of these compounds with established drug-likeness rules. While CMNPD8663 falls short on several parameters, CMNPD8904 is aligned with Lipinski\u0026apos;s Rule, indicating its potential as an orally available drug. Notably, CMNPD11566 stands out, conforming to all the applied rules, including Lipinski, Pfizer, GSK, and the Golden Triangle, marking it as the most promising candidate in terms of drug-likeness.\u003c/p\u003e\n\u003cp\u003eFocusing on CMNPD11566, which emerges as the front-runner in drug-likeness, its ADMET characteristics are explored in Supplementary Table 1. Although the Caco-2 permeability data suggest some absorption challenges, the MDCK permeability values offer a counterbalance, indicating the possibility of effective passive diffusion. The high plasma protein binding percentage points to a significant bound fraction in circulation, with only a small percentage available for active pharmacological effect. Metabolically, the compound engages with a variety of Cytochrome P450 enzymes, a factor that requires careful monitoring due to the potential for drug-drug interactions.\u003c/p\u003e\n\u003cp\u003eThe clearance and half-life data of CMNPD11566 imply a moderate rate of elimination and a brief systemic presence, respectively, influencing dosing intervals. Toxicity predictions, though varied, signal the need for careful consideration of the compound\u0026apos;s safety profile, especially in areas of genotoxic carcinogenicity and skin sensitization.\u003c/p\u003e\n\u003cp\u003eIn conclusion, while the physicochemical properties of CMNPD11566 provide a favorable starting point, and its adherence to drug-likeness rules is exemplary, the comprehensive ADMET profile underscores the need for a nuanced approach to its development. Thorough preclinical studies will be essential to navigate its complex ADMET landscape and fully elucidate its therapeutic potential.\u003c/p\u003e\n\u003cp\u003e2.4 Molecular Dynamics Simulation and Biophysical Analysis of CMNPD11566 and Its Derivative HCJ007 in Interaction with SQLE\u003c/p\u003e\n\u003cp\u003eMolecular dynamics simulations are essential in analyzing the structure-activity relationships of drug candidates, providing critical insights into their interactions with target proteins. In the 1000 ns molecular dynamics simulation analysis of CMNPD11566, as illustrated in Figures 4A-C, we identified several potential limitations. Figure 4A displays the RMSD trajectories of CMNPD11566, with fluctuations ranging between 1 to 3 \u0026Aring;. Although this indicates relatively stable interactions with the SQLE enzyme, there is still room for further optimization. Figure 4B shows that while CMNPD11566 has high-frequency interactions with residues like Tyr-195 and Thr-417, it also exhibits extremely low interaction frequencies with residues such as Met-196 and Ile-197. This suggests that the structure-activity relationship of CMNPD11566 with SQLE could potentially be compromised by its inherent instability. Further, Figure 4C reveals the reasons behind many low-frequency interaction sites of CMNPD11566; notably, its alkyl chain fails to form stable interactions with residues in the pocket.\u003c/p\u003e\n\u003cp\u003eGiven the mobility of CMNPD11566\u0026apos;s alkyl chain, which could lead to binding stability issues, we deemed structural modification necessary. By carefully analyzing the interaction patterns of CMNPD11566 and incorporating structural insights from known target inhibitors Naftifine and Butenafine, we designed a new compound, HCJ007. This approach aimed to address the identified binding issues of CMNPD11566 while preserving its effective interactions with key residues.\u003c/p\u003e\n\u003cp\u003eSubsequent simulations of HCJ007, as depicted in Figures 4D-F, showed significant improvements in its dynamic characteristics. Figure 4D reveals a more stable RMSD for HCJ007, indicating enhanced binding conformation stability. Figure 4E shows that HCJ007 has increased interaction frequencies with key residues, and significantly fewer low-frequency interaction sites. Figure 4F further illustrates the detailed interaction patterns of HCJ007, revealing stronger and more consistent binding interactions compared to CMNPD11566, suggesting that it may have more effective binding and therapeutic action.\u003c/p\u003e\n\u003cp\u003eIn orchestrating the development of HCJ007 as a potent therapeutic against SQLE, a series of intricate biophysical analyses were pivotal, each unraveling a layer of interaction dynamics and molecular stability. Throughout its modification process, HCJ007 consistently adhered to the four key drug-likeness rules, and notably, it displayed no PAINS (Pan-Assay Interference Compounds) alerts, as detailed in Supplementary Table 2. These insights are instrumental in sculpting HCJ007\u0026apos;s design, focusing on enhancing its specificity and efficacy.\u003c/p\u003e\n\u003cp\u003eInitially, Figure 5A reveals the evolving affinity between HCJ007 and SQLE through the lens of MMGBSA binding free energy. The trend of deepening negative values, averaging at -124.48 kcal/mol and bottoming out at -151.82 kcal/mol, illustrates an intensifying molecular interaction. This trend is not a mere data trajectory but underscores HCJ007\u0026apos;s potential as a high-affinity inhibitor, robustly engaging with SQLE.Building upon this overarching energy landscape, Figure 5B delves into the molecular intricacies, mapping the contribution of individual residues. The marked negative contributions of residues like Asn-193 and Asn-382 transcend being mere numerical values; they are pivotal in stabilizing the HCJ007-SQLE complex, guiding the precision in drug design and optimization.\u003c/p\u003e\n\u003cp\u003eTransitioning to the atomic scale, Figure 5C offers insights into HCJ007\u0026apos;s internal dynamics. Here, the RMSF increment starting from atom 16 suggests a deliberate retention of flexibility within the alkyl chain, essential for maintaining an adaptable binding conformation.This narrative of internal mobility and conformational adaptability is further complemented by Figure 5D\u0026apos;s torsion angle analysis. The minimal fluctuations in torsion angles lend credence to the structural integrity of the alkyl chain, ensuring its functional alignment during the binding interaction.\u003c/p\u003e\n\u003cp\u003eSegueing into the realm of molecular stability and interface dynamics, Figures 5E and 5F together illustrate a cohesive story. Figure 5E showcases the consistent pattern of intra-molecular hydrogen bonding, a hallmark of molecular stability and functional integrity. The steadiness punctuated by transient spikes mirrors the compound\u0026apos;s ability to dynamically respond to structural shifts within the complex, indicating a resilient yet adaptable binding profile. Complementing this, Figure 5F shifts the focus to the broader interface properties like SASA, MolSA, and PSA. Their concurrent trends, subtly modulating the hydrophobic-hydrophilic balance, narrate an intricate tale of the molecular interface. These forces are not merely coexisting; they are intricately intertwined, dictating the binding dynamics and stability of the molecular complex.\u003c/p\u003e\n\u003cp\u003eIn totality, these analytical vignettes paint a composite picture of HCJ007 as a meticulously tuned compound for efficacy. Each atomic and molecular interaction within HCJ007 is choreographed to ensure synergistic engagement with SQLE, setting the stage for HCJ007\u0026apos;s evolution as a promising therapeutic agent.\u003c/p\u003e"},{"header":"3. Discussion","content":"\u003cp\u003eLeveraging the diverse chemical landscape of marine-derived compounds, this study has identified promising squalene epoxidase (SQLE) inhibitors, advancing the search for effective pancreatic cancer treatments. The application of an active learning model has refined virtual screening processes, pinpointing compounds with high binding affinity and requisite stability within the SQLE active site. This innovative approach has enabled the discovery of compounds that not only exhibit high theoretical efficacy but also hold potential for significant advancements in SQLE inhibition.\u003c/p\u003e\n\u003cp\u003eBuilding on this foundation, our comprehensive analysis using multiple scoring criteria for the thirteen compounds has shown significant advantages in various aspects. Notably, the docking scores of all selected compounds exceeded -10, with CMNPD18677 achieving -11.644, indicating exceptionally strong binding affinity with SQLE\u003csup\u003e13\u003c/sup\u003e. This is further corroborated by the state penalty of zero for all compounds, emphasizing their high specificity in binding to SQLE and reducing the likelihood of non-specific interactions\u003csup\u003e14\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eAdditionally, ligand strain energy is a crucial metric for assessing how a ligand\u0026apos;s energy changes when binding to a target protein\u003csup\u003e15\u003c/sup\u003e. Ideally, lower ligand strain energy means that the ligand can more naturally adapt to the protein\u0026apos;s binding pocket, reducing the energy loss caused by conformational changes and increasing binding affinity. In our screening results, most selected compounds exhibited ligand strain energy values below 10 kcal/mol, significantly lower than known inhibitors like Butenafine (17.45 kcal/mol), Naftifine (9.666 kcal/mol), and Terbinafine (8.458 kcal/mol). For instance, compound 18775 had a strain energy of only 4.051 kcal/mol, much lower than Butenafine. This suggests that these compounds may bind more easily to SQLE, as they require less conformational change. These findings highlight the effectiveness of using an active learning model to select marine natural products as SQLE inhibitors. Compounds with low strain energy are theoretically more suitable as drug candidates, as they demand less energy to bind to their target proteins, potentially enhancing their efficacy and stability within the body.\u003c/p\u003e\n\u003cp\u003eMost impressively, the MMGBSA dG Bind scores reveal that all selected compounds have significantly lower scores compared to known inhibitors. Particularly remarkable is CMNPD18775, which scored -73.52 kcal/mol, far surpassing the scores of Butenafine, Naftifine, and Terbinafine. This indicates that the complexes formed with SQLE are likely to be more thermodynamically stable and potentially more effective as inhibitors\u003csup\u003e16\u003c/sup\u003e. This finding underscores the effectiveness of the active learning model in efficiently identifying potent and stable SQLE inhibitors from a vast repository of marine-derived compounds.\u003c/p\u003e\n\u003cp\u003eTransitioning to the dynamic characteristics, CMNPD11566 exhibited limitations in its interaction dynamics with SQLE, as seen in the molecular dynamics simulations. Its RMSD trajectories showed fluctuations within the range of 1-3 \u0026Aring;, suggesting areas for optimization despite relatively stable interactions\u003csup\u003e17\u003c/sup\u003e . The interaction frequency analysis revealed high-frequency interactions with residues like Tyr-195 and Thr-417 but lower interaction frequencies with others, indicating potential instability and suboptimal interaction with the enzyme\u0026apos;s binding pocket. Tyr-195 was identified as a crucial residue in the SQLE enzyme. This residue plays a significant role in inhibitor binding, as evidenced by the fact that both NB-598\u003csup\u003e18\u003c/sup\u003e and\u0026nbsp;Cmpd-4\u0026quot;\u0026nbsp;establish a hydrogen bond with Tyr-195, which is the only specific and directional interaction these compounds have with SQLE\u003csup\u003e19\u003c/sup\u003e. This interaction is consistent across all known SQLE inhibitors and explains the required presence of the tertiary amine motif in these inhibitors. The interaction with conserved Tyr-195 is critical for the binding and efficacy of SQLE inhibitors.\u003c/p\u003e\n\u003cp\u003eIn contrast, HCJ007 displayed markedly improved dynamic characteristics. Its more stable RMSD, enhanced binding conformation stability, and increased interaction frequencies with key residues, as seen in subsequent simulations, highlight its potential for more effective binding and therapeutic action. This comparison underscores the effectiveness of the structural modifications made to HCJ007, resulting in improved interaction dynamics essential for successful SQLE inhibition.\u003c/p\u003e\n\u003cp\u003eFinally, the ADMET data comparison reveals that HCJ007 offers significant improvements over CMNPD11566 in various parameters, such as lower BBB penetration and PPB, suggesting better central nervous system accessibility and higher active drug circulation\u003csup\u003e20\u003c/sup\u003e. Despite a higher clearance rate, HCJ007\u0026apos;s comparable half-life to CMNPD11566 suggests similar durations in the body. Additionally, HCJ007 shows significantly reduced toxicity in AMES test results and skin sensitization, lower risks of hERG cardiac toxicity and DILI, and potentially better drug-like properties as indicated by its higher QED score.\u003c/p\u003e\n\u003cp\u003eBased on these findings, the study propels us into new realms of research and clinical development, particularly focusing on the potential of HCJ007 as a SQLE inhibitor. With its superior dynamic characteristics and significant ADMET advantages over CMNPD11566, particularly in toxicity, safety, and drug-likeness, HCJ007 stands as a strong candidate for further exploration. This research lays a foundation for innovative approaches in drug discovery, combining computational modeling with empirical data to refine drug efficacy and safety profiles. The next steps involve validating these results through in vitro and in vivo studies, aiming towards eventual clinical trials. Such progression towards translational research emphasizes the crucial role of comprehensive preclinical evaluations, ensuring that compounds like HCJ007 are not only effective in laboratory settings but also well-suited for real-world therapeutic applications.\u003c/p\u003e"},{"header":"4. Materials and Methods","content":"\u003cp\u003e4.1\u0026nbsp;Protein Preparation\u003c/p\u003e\n\u003cp\u003eIn our quest to discover inhibitors targeting SQLE, we adopted structure-based computational techniques, focusing on the FGFR3 crystal structure (PDB ID: 6C6N) complexed with Cmpd-4\u0026quot;\u003csup\u003e19\u003c/sup\u003e. This particular structure was chosen to serve as the receptor model. The preparation of the receptor-ligand complex was carried out using Schr\u0026ouml;dinger\u0026rsquo;s Protein Preparation Wizard, which involved a series of meticulous steps. These included the addition of missing hydrogen atoms, adjustment of metal ion states, bond order determination in HET groups, assessment and optimization of ligand protonation states along with their energy implications, tuning protonation states of histidine residues, correcting any misplaced heavy atoms, refining the hydrogen bonding network within the protein, and executing a restrained minimization to ensure structural integrity.The identified binding site within the 3D structure of the receptor, where Cmpd-4\u0026quot; interacts, was designated as the focal point for screening potential ligands. Consequently, a grid corresponding to this target site was generated to facilitate the screening process.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e4.2 Active learning based virtual screening\u003c/p\u003e\n\u003cp\u003eActive Learning Glide will generate a receptor grid from a prepared protein, prepare the Comprehensive Marine Natural Products Database (https://cmnpd.org/), and dock a subset of these ligands using Glide SP\u003csup\u003e21\u003c/sup\u003e. Active Learning workflows train a machine learning (ML) model on physics-based data, such as FEP+\u003csup\u003e22\u003c/sup\u003e predicted affinities or Glide docking scores, iteratively sampled from a full library using Schr\u0026ouml;dinger\u0026apos;s deep-learning powered QSAR platform, DeepAutoQSAR (https://www.schrodinger.com/science-articles/benchmark-study-deepautoqsar-chemprop-and-deeppurpose-admet-subset-therapeutic-data).\u0026nbsp;3\u0026nbsp;iterative training rounds\u0026nbsp;were\u0026nbsp;set. After all the ligands have been screened using the last model, a selection of the top ligands will then be docked using Glide SP. The results of the docking were then quantified based on the consensus of docking scores and Prime MM-GBSA energy\u003csup\u003e23,24\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e4.3 Binding pose metadynamics\u003c/p\u003e\n\u003cp\u003eTo verify the stability of binding poses of the selected ligands from molecular dynamics (MD) simulations in both binding sites, a sequence of metadynamics MD simulations was executed, each lasting for 10 ns, on various docked poses and MD-established stable protein-ligand complexes\u003csup\u003e25\u0026ndash;27\u003c/sup\u003e. The chosen collective variable (CV) was the root mean square deviation (RMSD) of the ligand\u0026apos;s heavy atoms from their initial positions, computed post alignment of the binding sites to mitigate any drift. The parameters for the hill\u0026apos;s height and width in the metadynamics simulations were set at 0.05 kcal/mol and 0.02 \u0026Aring;, respectively. The system underwent solvation in a box with a 10 \u0026Aring; buffer, followed by a series of minimization steps, gradually elevating the system\u0026apos;s temperature to 300 K and alleviating any initial structural stresses or contacts. The evaluation of stability was based on monitoring the RMSD fluctuations of the ligand throughout the simulation (termed as PoseScore) and quantifying the average duration of critical contacts between the ligand and protein residues (referred to as PersScore).\u003c/p\u003e\n\u003cp\u003e4.4 ADMET Screening and Drug-Likeness Predictions\u003c/p\u003e\n\u003cp\u003eTo ensure favorable ADMET (absorption, distribution, metabolism, excretion, and toxicity) properties and non-toxicity of potential drug candidates, the study employed ADMETLAB 2.0 (https://admetmesh.scbdd.com/)\u003csup\u003e28\u003c/sup\u003e for prediction. Adherence to medicinal chemistry guidelines, such as the Lipinski\u003csup\u003e29\u003c/sup\u003e, Pfizer\u003csup\u003e30\u003c/sup\u003e, GSK\u003csup\u003e31\u003c/sup\u003e, and Golden Triangle rules\u003csup\u003e32\u003c/sup\u003e, was pivotal in identifying compounds with optimal ADMET attributes.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e4.5 Molecular dynamic simulation\u003c/p\u003e\n\u003cp\u003eThe Desmond software from Schr\u0026ouml;dinger LLC was employed for conducting molecular dynamics (MD) simulations over a period of 1000 ns. The TIP3P water model, representing a three-point intermolecular interaction potential, was chosen for these simulations. The simulations were set up within an orthorhombic box, maintaining a constant temperature of 300 K and pressure of 1 atm, while utilizing the OPLS 2005 force field\u003csup\u003e33\u003c/sup\u003e. To achieve neutrality in the models and mimic physiological conditions, counter ions were added, and a 0.15 M sodium chloride solution was used to simulate ligand-binding status in physiological environment\u003csup\u003e34\u003c/sup\u003e. Prior to the commencement of the simulations, the models underwent a relaxation phase, and during the simulations, data was recorded and stored every 100 ps for later analysis.\u003c/p\u003e"},{"header":"5. Conclusions","content":"\u003cp\u003eThis study marks a significant advancement in identifying potent squalene epoxidase (SQLE) inhibitors for pancreatic cancer treatment. Utilizing an active learning model, we identified compounds, notably HCJ007, with superior binding affinity, stability, and ADMET properties compared to known inhibitors. HCJ007\u0026apos;s enhanced dynamic characteristics and reduced toxicity profiles position it as a promising therapeutic candidate. These findings pave the way for further validation through in vitro and in vivo studies, potentially leading to clinical trials and offering new avenues in cancer treatment.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eData availability:\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003eAll data generated or analysed during this study are included in this published article [and its supplementary information files].\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e8.Author Contributions:\u003c/strong\u003e Conceptualization, YiPing Mou; Data curation, Yunyun Xu and YouJian Xu; Funding acquisition, YiPing Mou; Project administration, YiPing Mou; Resources, Qiang Wang; Validation, Yunyun Xu; Writing \u0026ndash; original draft, Yunyun Xu; Writing \u0026ndash; review \u0026amp; editing, YiPing Mou. All authors read and approved the final manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e9. Funding:\u003c/strong\u003e This research was funded by Scientific Research Fund of Zhejiang Provincial Education Department, grant number Y202249377.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflicts of Interest:\u003c/strong\u003e The authors declare no conflict of interest.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eHalbrook, C. J., Lyssiotis, C. A., Pasca di Magliano, M. \u0026amp; Maitra, A. Pancreatic cancer: Advances and challenges. Cell 186, 1729\u0026ndash;1754 (2023).\u003c/li\u003e\n\u003cli\u003eXiao, M. et al. Functional significance of cholesterol metabolism in cancer: from threat to treatment. Exp Mol Med 55, 1982\u0026ndash;1995 (2023).\u003c/li\u003e\n\u003cli\u003eYou, W. et al. SQLE, A Key Enzyme in Cholesterol Metabolism, Correlates With Tumor Immune Infiltration and Immunotherapy Outcome of Pancreatic Adenocarcinoma. Front Immunol 13, 864244 (2022).\u003c/li\u003e\n\u003cli\u003eXu, R. et al. SQLE promotes pancreatic cancer growth by attenuating ER stress and activating lipid rafts-regulated Src/PI3K/Akt signaling pathway. Cell Death Dis 14, 497 (2023).\u003c/li\u003e\n\u003cli\u003eKhalifa, S. A. M. et al. Marine Natural Products: A Source of Novel Anticancer Drugs. Mar Drugs 17, 491 (2019).\u003c/li\u003e\n\u003cli\u003eFrediansyah, A. The antiviral activity of iota-, kappa-, and lambda-carrageenan against COVID-19: A critical review. Clin Epidemiol Glob Health 12, 100826 (2021).\u003c/li\u003e\n\u003cli\u003eCarter, N. J. \u0026amp; Keam, S. J. Trabectedin: a review of its use in soft tissue sarcoma and ovarian cancer. Drugs 70, 355\u0026ndash;376 (2010).\u003c/li\u003e\n\u003cli\u003eDonato, E. M., Fern\u0026aacute;ndez-Zarzoso, M., Hueso, J. A. \u0026amp; de la Rubia, J. Brentuximab vedotin in Hodgkin lymphoma and anaplastic large-cell lymphoma: an evidence-based review. Onco Targets Ther 11, 4583\u0026ndash;4590 (2018).\u003c/li\u003e\n\u003cli\u003eBanik, R. K. \u0026amp; Engle, M. P. Ziconotide for Management of Cancer Pain Refractory to Pharmacotherapy: An Update. Pain Med 21, 3253\u0026ndash;3259 (2020).\u003c/li\u003e\n\u003cli\u003eMcConnell, O. J., Longley, R. E. \u0026amp; Koehn, F. E. The discovery of marine natural products with therapeutic potential. Biotechnology 26, 109\u0026ndash;174 (1994).\u003c/li\u003e\n\u003cli\u003eAbduljalil, J. M., Elfiky, A. A., Sayed, E.-S. T. A. \u0026amp; AlKhazindar, M. M. Computational identification of drug-like marine natural products as potential RNA polymerase inhibitors against Nipah virus. Comput Biol Chem 104, 107850 (2023).\u003c/li\u003e\n\u003cli\u003eTran, D. P. et al. Using molecular dynamics simulations to prioritize and understand AI-generated cell penetrating peptides. Sci Rep 11, 10630 (2021).\u003c/li\u003e\n\u003cli\u003eMeli, R., Morris, G. M. \u0026amp; Biggin, P. C. Scoring Functions for Protein-Ligand Binding Affinity Prediction using Structure-Based Deep Learning: A Review. Front Bioinform 2, 885983 (2022).\u003c/li\u003e\n\u003cli\u003eFriesner, R. A. et al. Extra precision glide: docking and scoring incorporating a model of hydrophobic enclosure for protein-ligand complexes. J Med Chem 49, 6177\u0026ndash;6196 (2006).\u003c/li\u003e\n\u003cli\u003eTong, J. \u0026amp; Zhao, S. Large-Scale Analysis of Bioactive Ligand Conformational Strain Energy by Ab Initio Calculation. J Chem Inf Model 61, 1180\u0026ndash;1192 (2021).\u003c/li\u003e\n\u003cli\u003ePattar, S., Adhoni, S. A., Kamanavalli, C. M. \u0026amp; Kumbar, S. S. In silico molecular docking studies and MM/GBSA analysis of coumarin-carbonodithioate hybrid derivatives divulge the anticancer potential against breast cancer. Beni-Suef University Journal of Basic and Applied Sciences 9, (2020).\u003c/li\u003e\n\u003cli\u003eKumar, M. et al. Definition of fatty acid transport protein-2 (FATP2) structure facilitates identification of small molecule inhibitors for the treatment of diabetic complications. Int J Biol Macromol 244, 125328 (2023).\u003c/li\u003e\n\u003cli\u003eHorie, M. et al. NB-598: a potent competitive inhibitor of squalene epoxidase. J Biol Chem 265, 18075\u0026ndash;18078 (1990).\u003c/li\u003e\n\u003cli\u003ePadyana, A. K. et al. Structure and inhibition mechanism of the catalytic domain of human squalene epoxidase. Nat Commun 10, 97 (2019).\u003c/li\u003e\n\u003cli\u003eTong, X. et al. Blood-brain barrier penetration prediction enhanced by uncertainty estimation. J Cheminform 14, 44 (2022).\u003c/li\u003e\n\u003cli\u003eFriesner, R. A. et al. Glide: a new approach for rapid, accurate docking and scoring. 1. Method and assessment of docking accuracy. J Med Chem 47, 1739\u0026ndash;1749 (2004).\u003c/li\u003e\n\u003cli\u003eWang, L. et al. Accurate and reliable prediction of relative ligand binding potency in prospective drug discovery by way of a modern free-energy calculation protocol and force field. J Am Chem Soc 137, 2695\u0026ndash;2703 (2015).\u003c/li\u003e\n\u003cli\u003eMittal, L., Kumari, A., Suri, C., Bhattacharya, S. \u0026amp; Asthana, S. Insights into structural dynamics of allosteric binding sites in HCV RNA-dependent RNA polymerase. J Biomol Struct Dyn 38, 1612\u0026ndash;1625 (2020).\u003c/li\u003e\n\u003cli\u003eMittal, L., Kumari, A., Srivastava, M., Singh, M. \u0026amp; Asthana, S. Identification of potential molecules against COVID-19 main protease through structure-guided virtual screening approach. J Biomol Struct Dyn 39, 3662\u0026ndash;3680 (2021).\u003c/li\u003e\n\u003cli\u003eClark, A. J. et al. Prediction of Protein-Ligand Binding Poses via a Combination of Induced Fit Docking and Metadynamics Simulations. J Chem Theory Comput 12, 2990\u0026ndash;2998 (2016).\u003c/li\u003e\n\u003cli\u003eFusani, L., Palmer, D. S., Somers, D. O. \u0026amp; Wall, I. D. Exploring Ligand Stability in Protein Crystal Structures Using Binding Pose Metadynamics. J Chem Inf Model 60, 1528\u0026ndash;1539 (2020).\u003c/li\u003e\n\u003cli\u003ePurushotham, N. et al. Design and synthesis of amino acid derivatives of substituted benzimidazoles and pyrazoles as Sirt1 inhibitors. RSC Adv 12, 3809\u0026ndash;3827 (2022).\u003c/li\u003e\n\u003cli\u003eXiong, G. et al. ADMETlab 2.0: an integrated online platform for accurate and comprehensive predictions of ADMET properties. Nucleic Acids Res 49, W5\u0026ndash;W14 (2021).\u003c/li\u003e\n\u003cli\u003eLipinski, C. A., Lombardo, F., Dominy, B. W. \u0026amp; Feeney, P. J. Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. Adv Drug Deliv Rev 46, 3\u0026ndash;26 (2001).\u003c/li\u003e\n\u003cli\u003eHughes, J. D. et al. Physiochemical drug properties associated with in vivo toxicological outcomes. Bioorg Med Chem Lett 18, 4872\u0026ndash;4875 (2008).\u003c/li\u003e\n\u003cli\u003eGleeson, M. P. Generation of a set of simple, interpretable ADMET rules of thumb. J Med Chem 51, 817\u0026ndash;834 (2008).\u003c/li\u003e\n\u003cli\u003eJohnson, T. W., Dress, K. R. \u0026amp; Edwards, M. Using the Golden Triangle to optimize clearance and oral absorption. Bioorg Med Chem Lett 19, 5560\u0026ndash;5564 (2009).\u003c/li\u003e\n\u003cli\u003eShivakumar, D. et al. Prediction of Absolute Solvation Free Energies using Molecular Dynamics Free Energy Perturbation and the OPLS Force Field. J Chem Theory Comput 6, 1509\u0026ndash;1519 (2010).\u003c/li\u003e\n\u003cli\u003eRasheed, M. A. et al. Identification of Lead Compounds against Scm (fms10) in Enterococcus faecium Using Computer Aided Drug Designing. Life (Basel) 11, 77 (2021).\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Table","content":"\u003cp\u003eTable 1 is available in the Supplementary Files section.\u003c/p\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":"Marine Natural Products, Screening of SQLE Inhibitors, Active Learning Model, ADMET Analysis, Molecular Dynamics Simulations, Molecular Modification","lastPublishedDoi":"10.21203/rs.3.rs-4163089/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4163089/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"In this study, we leveraged a sophisticated active learning model to enhance virtual screening for SQLE inhibitors. The model's improved predictive accuracy identified compounds with significant advantages in binding affinity and thermodynamic stability. Detailed analyses, including molecular dynamics simulations and ADMET profiling, were conducted, particularly focusing on compounds CMNPD11566 and its derivative HCJ007. CMNPD11566 showed stable interactions with SQLE, but HCJ007 exhibited enhanced dynamic characteristics and interaction patterns. ADMET data comparison highlighted HCJ007's superior profile in terms of lower toxicity and better drug-likeness. Our findings suggest HCJ007 as a promising candidate for SQLE inhibition, with significant improvements over CMNPD11566 in various pharmacokinetic and safety parameters. The study underscores the efficacy of computational models in drug discovery and the importance of comprehensive preclinical evaluations.","manuscriptTitle":"Active Learning-Driven Discovery and Dynamics Simulation of Novel SQLE Inhibitors with ADMET Analysis and Molecular Modification","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-04-16 07:56:47","doi":"10.21203/rs.3.rs-4163089/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":"845a5d1a-9427-4a18-91ac-049696e239c4","owner":[],"postedDate":"April 16th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":30713250,"name":"Biological sciences/Cancer/Cancer prevention"},{"id":30713251,"name":"Biological sciences/Cancer/Cancer therapy"}],"tags":[],"updatedAt":"2025-01-03T10:24:05+00:00","versionOfRecord":[],"versionCreatedAt":"2024-04-16 07:56:47","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-4163089","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4163089","identity":"rs-4163089","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}