Structural Prerequisites for Pim-1 Kinase Inhibition and the Identification of a Pyrazolyl-Methylene-Indolinone Hit Authors: Rufaida Al Zoubi, Omar Hadieh, Soraya Alnabulsi, Israa H Isawi, Shubin Liu Published: 2026 ABSTRACT AbstractIntroductionPim-1 is a serine/threonine kinase of the pan-Pim family (Pim-1, −2, and −3), whose overexpression is associated with various cancers. Its distinctive hinge-region architecture provides opportunities for selective inhibitor development, with several inhibitors reported as clinical candidates. Understanding structure–activity relationships is essential for distinguishing biologically active compounds from high-scoring docking hits, a major challenge in structure-based drug discovery.MethodsA virtual screening workflow comprising Glide docking and Induced Fit Docking (IFD) was applied to approximately 900,000 commercially available compounds. Six structurally diverse compounds were selected for in vitro Pim-1 inhibitory activity evaluation. Molecular dynamics (MD) simulations and quantum mechanical (QM) calculations were subsequently performed to investigate the structural determinants of inhibitory activity.ResultsAmong the tested compounds, compound 6, representing a previously unreported pyrazolyl-methylene-indolinone scaffold for Pim-1 inhibition, was identified as a Pim-1 inhibitory hit with an IC50 value of 5.77 µM. QM calculations revealed differences in the energetic accessibility of coplanar geometries between the active and inactive compounds, providing a structural explanation for the discrepancy between docking predictions and experimental activity. Moreover, MD simulations of Pim-1 in complex with compound 6 revealed the formation of dual polar interactions between its pyrazole ring and catalytic-region residues K67 and E89.DiscussionThese findings provide insight into the structural and conformational requirements for productive Pim-1 inhibition and highlight the value of integrating docking, MD simulations, and QM calculations during the search for novel inhibitors. The identified hit represents a previously unreported scaffold for Pim-1 inhibition that may provide a starting point for future optimization.Keywords: pim-1 kinase, structure-activity relationship, 1H-pyrazolyl-methylene-indolinone, virtual screening, quantum mechanical calculations, molecular dynamics simulation Introduction The Pim-1 kinase is a proto-oncogene serine/threonine kinase that belongs to the pan-Pim family, comprising Pim-1, Pim-2, and Pim-3.1 Pim-1 plays a vital role in cellular processes such as survival, differentiation, proliferation, and apoptosis. The constitutively active enzyme is encoded by the pim-1 gene, which was first described in murine T-cell lymphomas and subsequently reported to show abnormal overexpression in other hematologic and non-hematologic malignancies, including B-cell non-Hodgkin’s lymphoma, breast cancer, prostate cancer, and acute myeloid leukemia.2,3 Unlike many kinases, Pim kinases lack a regulatory domain, rendering them constitutively active; thus, their enzymatic activity is predominantly regulated at the transcriptional and translational levels.4,5 This persistent activation has been implicated in uncontrolled tumor growth, decreased overall survival, and the development of drug resistance in various cancers, thereby underpinning the rationale for co-targeting Pim kinases alongside existing therapeutic strategies to enhance treatment efficacy.2 Indeed, clinical evaluation of Pim inhibitors has provided support for the therapeutic potential of Pim inhibition; nevertheless, no Pim-targeted inhibitor has yet been approved, and the development of some clinical candidates has been limited by off-target toxicity and pharmacological liabilities. Clinical candidates such as SGI-1776, AZD-1208, PIM447, and CX-6258, each representing a different chemical scaffold, exemplify the substantial efforts directed toward pharmacological inhibition of Pim kinases (Table 1).6 Table 1A Selected Set of Pim-1 Inhibitors Showing Their Chemical Scaffold and Highlighting Their Interacting MoietiesCompound IDChemical ScaffoldChemical StructureIn vitro [1]PDB IDSGI-1776Imidazo[1,2-b]pyridazines2D skeletal structure with piperidine, N rich fused rings and phenoxy trifluoromethyl substituent.IC50 (7 nM)4A7CCX-6258Oxindole2D skeletal formula of a multi ring compound with piperazine, chloro and carbonyl groups.IC50 (5 nM)5O13AZD1028Thiazolidine-2,4-dioneTwo dimensional skeletal structure with thiazole, piperidine, fused bicyclic aromatic ring and NH2 groups.IC50 (3 nM)NAPIM447PyridineaminesTwo-dimensional skeletal formula of a fluorinated fused ring linked by an amide to two N rings.Ki (6 pM)5DWRGDC-0339Diaminopyrazole2D skeletal formula of a multi ring molecule with fluorinated benzene, sulfur link, amide and amines.Ki (0.03 nM)6NO9Notes: Color coding cyan: catalytic site-interacting moiety, green: hinge-region interacting moiety, and magenta: acidic patch interacting moiety. A Selected Set of Pim-1 Inhibitors Showing Their Chemical Scaffold and Highlighting Their Interacting Moieties Notes: Color coding cyan: catalytic site-interacting moiety, green: hinge-region interacting moiety, and magenta: acidic patch interacting moiety. Structurally, Pim kinases are characterized by a distinctive hinge-region motif, ERP123XP125, differentiating them from other kinase families. This motif, comprising two proline residues, drastically influences the geometry of the adenine-binding region, disrupting one of the conserved hydrogen-bond interactions typically observed between ATP and the kinase hinge-region backbone (Figure 1).4,7 Such unique structural features of Pim kinases have prompted intensive research toward developing selective small-molecule inhibitors, exploiting this distinction to achieve beneficial selectivity and potential anti-cancer efficacy.8,9 In addition, structural characterization of Pim-1 has significantly advanced, and multiple structures of Pim-1 complexed with different chemical scaffolds of various inhibitors have been deposited in the Protein Data Bank (PDB) (https://www.rcsb.org/),10 offering critical insights into structure activity relationships (SAR) that would guide inhibitor development. (A) Structural alignment of different serine/threonine kinases with the ERPxP motif in Pim kinases. (B) Overlay of PIM1 (PDB ID: 3A99) on CDK2 (PDB ID: 8FP5) crystal structures showing different hinge region conformation. (C) Hinge region’s interaction with ATP in Pim-1 (PDB ID: 5O116) shows the lack of H-bond interaction with Pro123 compared to (D) dual H-bond interaction of the ATP with the hinge region in CDK2. Accordingly, recent Pim-1 inhibitor discovery studies have increasingly employed integrated computational approaches, including virtual screening, molecular docking, machine learning, and molecular dynamics simulations, to identify and characterize new inhibitor scaffolds. However, a comprehensive understanding of structure–activity relationships remains essential in drug discovery, as predictive algorithms and docking scores alone may not fully capture target-specific structural and conformational features that are critical for ligand activity, emphasizing the importance of complementary approaches for evaluating docking-derived hits.11–14 In our study aimed at identifying new Pim-1 kinase inhibitors, virtual screening of commercially available chemical databases was conducted to identify chemical scaffolds not previously associated with Pim-1 inhibitory activity. Following successive docking studies and detailed visual inspection of predicted binding interactions within Pim-1’s active site, six compounds were selected for in vitro Pim-1 inhibitory activity evaluation. Subsequent computational quantum mechanics (QM) calculations and molecular dynamics (MD) simulations were employed to evaluate the binding pose of the active compound and to rationalize SAR comprehensively. The QM analysis provided insight into the conformational requirements associated with Pim-1 inhibition, emphasizing the energetic accessibility of co-planarity between the catalytic lysine-binding and hinge-binding moieties as a potential structural prerequisite for Pim-1 inhibitory activity. Among the investigated compounds, the pyrazol-4-yl-methylene-indolin-2-one derivative (compound 6), which favored this co-planarity, displayed Pim-1 kinase inhibitory activity with an IC50 value of 5.77 µM. Additionally, MD simulation of Pim-1 in complex with compound 6 revealed distinct polar interactions within the Pim-1 catalytic region. Collectively, these findings offer structural insights that advance the understanding of effective Pim-1 kinase inhibition, while the identified pyrazolyl-methylene-indolinone hit may provide a starting point for further optimization toward more potent Pim-1 inhibitors. Methods Computational Methods Protein Selection and Preparation The X-ray structure of human Pim-1 kinase (PDB ID: 5DWR; resolution at 2.00 Å) was retrieved from the Protein Data Bank.15 Using Schrödinger Suite 2022–4, the Protein Preparation Wizard assigned protonation state using PROPKA at pH 7.0,16 added hydrogen atoms, and minimized hydrogen under default settings.17 Virtual Screening and Docking Studies A library of approximately 900,000 compounds retrieved from the Maybridge HitDiscover, ASINEX, and National Cancer Institute (NCI) databases was prepared in LigPrep (Schrödinger). Ionization states were predicted using Epik at pH 7 ± 2, and a maximum of 32 possible stereoisomers were generated per ligand.18 The prepared compound library was initially subjected to a hierarchical docking workflow comprising Glide Standard Precision (SP) docking followed by Induced Fit Docking (IFD), as detailed below. Prior to virtual screening, both docking protocols were validated by re-docking the co-crystallized ligand into the ATP-binding site using the same docking parameters applied during library screening. The ability of the docking protocols to reproduce the experimental binding mode was assessed by calculating the heavy-atom root-mean-square deviation (RMSD) between the crystallographic and re-docked ligand poses. Glide Protein Grid Generation was defined using the co-crystallized inhibitor, with van der Waals radii scaled to 0.9 for all binding site atoms. Standard Precision Glide screening was initially performed using Ligand Docking (Schrödinger) with the following parameters: flexible ligand sampling, ligand van der Waals radii scaled to 0.9 for all atoms, Epik state penalties applied to docking scores, and the maximum number of reported poses set to one per ligand. Compounds were ranked according to their Glide SP docking scores and subsequently inspected for their ability to establish key interactions with residues known to govern Pim-1 molecular recognition.19,20 No predefined docking-score cutoff, drug-likeness criteria, or PAINS filters were applied during this selection process. Priority was given to compounds capable of interacting with the catalytic lysine (K67), while additional interactions involving the hinge region, G-loop, and acidic patch (D128, D131, and E171) were considered during visual inspection. This selection strategy reduced the initial screening library to fourteen structurally diverse candidates, whose binding poses were subsequently re-evaluated using Induced Fit Docking (IFD). The docking box in IFD was centered on the co-crystallized inhibitor. A standard IFD protocol was employed for ligand conformational sampling, while both protein and ligand van der Waals scaling factors were set to 1.0, and residues within 7 Å of each ligand pose were allowed to undergo structural refinement.21 Accordingly, six compounds were selected for biochemical evaluation based on their predicted binding modes involving key binding interactions, and scaffold diversity. Quantum Mechanical Calculations Interaction Energy Calculations Density functional theory (DFT) calculations were carried out using the Gaussian 16 C.01 software package.22 Two models representing the active and inactive complexes, consisting of 511 and 475 atoms, respectively, were first fully optimized using the ONIOM model.23 The substrate was treated at the M062X/6-311+G(d)24 level of theory, while the protein residues were modeled at the M062X/6-31G(d)25 level. The fully optimized structures were subsequently used to calculate the basis set superposition error (BSSE)-corrected interaction energies between the substrate and the protein at the M062X/6-311+G(d) level of theory.26 Electrostatic Potential Maps and Conformational Analysis QM calculations were done using the Spartan20 software, WaveFunction, Inc. Electrostatic potential maps of selected ligands were determined in water using the Hartree-Fock method with the 6–31G* basis set (HF/6-31G*). To determine global minima and perform force field parameterization, including dihedral energy scans, calculations were conducted in the gas phase using second-order Møller-Plesset perturbation theory (MP2) with the 6–31* polarization basis set (MP2/6-31G*). Molecular Dynamics Simulations MD simulations were conducted for the free Pim-1 protein and its complexes with compounds 5 and 6, obtained from IFD. CHARMM-GUI Input Generator was used to generate simulation cells, including solvating the protein complexes in an octahedral TIP3P water box and neutralizing it with 0.15 M NaCl and generating Amber-compatible input files.27,28 Initial CHARMM force field parameters for the ligands were retrieved from the ParamChem CGenFF interface,29,30 with high-penalty parameters further optimized by fitting to MP2-generated energy profiles.31 Particle mesh Ewald periodic boundary conditions and the CHARMM36m force field with the WYF parameter for cation-π interactions were used.32 Equilibration was carried out under NVT conditions for 100 ps, followed by a 500 ns production run under NPT conditions using Amber21 tools.33 Trajectory analysis was performed in VMD.34 RMSD calculations for the active site were based on residues 44, 45, 46, 49, 52, 67, 104, 120, 121, 122, 123, 126, 128, 131, 185, and 186. Experimental Methods Pim-1 Kinase Inhibitory Activity Evaluation The six compounds selected for biological evaluation were purchased from the Maybridge HitDiscover collection (Thermo Fisher Scientific, Waltham, MA, USA), assigned internal identifiers (Compounds 1–6), and used as received. Their corresponding commercial Maybridge compound identifiers are provided in Table S1 (Supplementary). The inhibitory activity of Compounds 1–6 against human Pim-1 kinase was evaluated in vitro using the Z′-LYTE fluorescence-based biochemical assay performed by Thermo Fisher Scientific, USA. Assay Protocol and Data Analysis Each test compound (1–6) was prepared as a 100% DMSO solution and dispensed at 100 nL per well to yield final assay concentrations of 100.0, 33.3, 11.1, 3.7, 1.23, 0.412, 0.137, 0.0457, 0.0152, and 0.00508 µM. To each well, 5 µL of a reagent mixture containing Pim-1 kinase (final 0.3–1.19 ng) and the Ser/Thr 07 substrate peptide (final concentration: 2.0 µM; sequence incorporating a PKC consensus motif: (R⁄K)_1–3(X)_0–2(S⁄T)(X)_0–2(R⁄K)_1–3, where X is any residue) was added. Subsequently, 2.5 µL of ATP solution (ATP was added at its apparent Km concentration (400 µM)), and 2.4 µL of kinase buffer (50 mM HEPES, pH 7.5; 0.01% BRIJ-35; 10 mM MgCl2; 1 mM EGTA) were added to each well. The mixture was shaken for 30 seconds and then incubated for 1.0 hr at RT before 5 µL of a 1:45000 dilution of the development reagent was added. After 1 hr incubation at RT, the fluorescence signals were measured at an excitation wavelength of 400 nm and an emission wavelength of 445 nm (coumarin emission) and 520 nm (fluorescein emission). Appropriate assay controls were included according to the standard SelectScreen™ Z′-LYTE™ protocol, including phosphorylation and inhibition controls, compound-interference controls, and the PIM1 reference inhibitor TBB (IC50 value of 1120 nM). Dose-response curves were analyzed by ThermoFisher Scientific using XLfit (IDBS), and IC50 values were determined by fitting the data to the XLfit Model 205 sigmoidal dose-response regression model. (Further details are provided in the Supplementary). Results and Discussion Initial Selection of the Pim-1 Crystal Structure A comprehensive set of available Pim-1 crystal structures was retrieved from the PDB and aligned using the Protein Structure Alignment protocol in Schrödinger Suite 2022–4. Based on visual inspection of the active site, the glycine-rich loop (G-loop) has been resolved in two distinct conformations; an open conformation, in which F49 faces away from the catalytic pocket, and a closed conformation, in which F49 is oriented toward the active site. In the closed conformation, the G-loop adopts a folded position over the ATP-binding site, thereby stabilizing the bound ligand at position. These conformational states are exemplified by Pim-1 in complex with ATP (PDB ID: 3A99)35 for the open conformation and Pim-1 in complex with PIM447 (PDB ID: 5DWR15) for the closed one (Figure 2A and B). While F49 projects into the catalytic site in the closed conformation, the loop near G45 creates a wider space (Figure 2). This configuration allows for better accommodation of structurally diverse ligands while maintaining hydrophobic interaction with F49 (Figure 2C). Consequently, the closed conformation (PDB ID: 5DWR) was selected for the subsequent virtual screening of the chemical libraries, as it provides a more enclosed binding environment that favors stabilizing interactions with the G-loop. G-loop conformation differences between the open (Pim-1 in complex with ATP, PDB ID: 3A99) (A) and closed (Pim-1 in complex with PIM447, PDB ID: 5DWR) (B) G-loop conformations, and an overlay (C) of the two conformations highlighting the differences in the conformation near the G45 and F49 residues within the G-loop affecting the binding of ligands to the active site of the enzyme. Computational Screening The chemical library databases, comprising around 900,000 compounds, were successfully prepared and filtered for compounds’ potential binding to the Pim-1 catalytic site using a docking workflow that comprises Glide SP docking, followed by IFD docking. Prior to compound library screening, the docking protocols were validated by re-docking the co-crystallized ligand into the Pim-1 ATP-binding site. Glide SP and Induced Fit Docking (IFD) successfully reproduced the crystallographic binding pose, with heavy-atom RMSD values of 0.27 Å and 1.00 Å, respectively as shown in Figure S1 (Supplementary), thereby confirming the suitability of both docking protocols for subsequent structure-based virtual screening. During the initial Glide SP screening, the van der Waals spheres for both the binding site atoms and the screened ligands were scaled to 90%. This was done to account for the rigidity of the protein structure inherent in the Glide algorithm, thereby increasing the tolerance for potential steric clashes. A key determinant interaction that has been reported for potent Pim-1 inhibition is a polar interaction with the catalytic lysine residue (K67). Additional structural features apparent in multiple Pim-1 crystal structures include hydrophobic interaction with gatekeeper and G-loop residues (including F49), as well as polar interaction with the backbone of E121 in the hinge region. Moreover, polar interactions with the acidic patch (D128, D131, and E171) in the C-terminal lobe have been associated with improving binding affinity (Table 1 and Figure 3).36 Pim-1 catalytic site. Interacting groups involved in potent inhibition are displayed as spheres; polar interaction with K67 at the catalytic site (Cyan), polar interaction with E121 in the hinge region (Green), polar interactions with the acidic patch (D128, D131, and E171) (Magenta), and ligands also engage in hydrophobic interaction with G-loop residues including F49 (Purple). Accordingly, the highest-ranking Glide SP hits were prioritized based on their ability to engage with the catalytic K67 and then visually inspected for additional favorable interactions. A set of fourteen structurally diverse compounds were selected for further investigation using IFD, in which the protein residues within 7 Å of the ligand were allowed to relax in order to account for protein flexibility. Following visual inspection of the refined docking poses, six compounds representing four structurally distinct chemical scaffolds were selected for subsequent in vitro evaluation (Table 2). Table 2Results from Computational Docking and in vitro Testing for Six Compounds Retrieved from a Commercially Available Chemical LibraryCompound IDChemical StructureInitial Docking Score (Kcal/mol)Induced-FitDocking Score(Kcal/mol)IC50 values (µM)112D skeletal formula of a substituted urea linking two benzene rings with cyano, trifluoromethyl and chlorine.−8.5−9.2>10022D skeletal formula of a fused ring system with benzene, multiple nitrogens and a hydroxyethylamino side chain.−8.6−8.7>1003Skeletal formula of a heterocyclic compound with benzene, pyridine and thiazole rings.−9.8−9.8>10042D skeletal structure with two benzene rings, one five-membered N,S ring and four N atoms.−8.4−8.4>10052D skeletal formula of a cyclohexane linked to a N and S containing ring and a benzene ring.−9.4−10.4>1006Skeletal chemical structure of a three-ring compound with O, N, NH and one C double bond O.−8.8−9.55.770Notes: 1The IC50 values were measured at least four times for each compound. Color coding resembles potential interactions identified from Induced Fit docking results; cyan: K67-interacting moiety, green: hinge-region interacting moiety, magenta: acidic patch interacting moiety, purple: G-loop interacting moiety. Results from Computational Docking and in vitro Testing for Six Compounds Retrieved from a Commercially Available Chemical Library Notes: 1The IC50 values were measured at least four times for each compound. Color coding resembles potential interactions identified from Induced Fit docking results; cyan: K67-interacting moiety, green: hinge-region interacting moiety, magenta: acidic patch interacting moiety, purple: G-loop interacting moiety. Analysis of the docking poses revealed that, with the exception of compound 1, all selected compounds occupied the hinge region of the ATP-binding site. Compounds 2–5 could form additional ionic interactions with the acidic patch, while compound 6 could form H-bond interactions with hinge region E121. Compounds 1, 2, and 6 exhibited π-π stacking interactions with G-loop residue F49. Notably, compounds 3 and 5, which possess a thienopyrimidine moiety capable of interacting with K67, were anticipated to exhibit favorable inhibitory activity based on their ability to establish multiple interactions within the active site, involving ionic interactions with the acidic patch, and due to their favorable docking scores (Table 2). Pim-1 Inhibitory Activity Evaluation The inhibitory activity of the six selected compounds against human Pim-1 kinase was evaluated using the Z′-LYTE fluorescence-based biochemical assay. As summarized in Table 2, compound 6 was the only compound to exhibit measurable inhibitory activity with an IC50 value of 5.77 µM. In contrast, the remaining compounds showed no measurable inhibition under the experimental conditions despite several of them exhibiting favorable docking scores. Notably, compound 5, which exhibited favorable docking profiles, had an IC50 >100 µM, underscoring the discrepancy between docking-based predictions and experimental activity. These findings highlight the limitations of docking scores alone in predicting biological activity and emphasize the importance of integrating experimental validation with molecular dynamics simulations and quantum mechanical calculations. To investigate the structural basis underlying the observed discrepancy between computational predictions and experimental activity, compounds 5 and 6 were selected for subsequent molecular dynamics simulations and quantum mechanical analyses. Molecular Dynamics Simulation To assess the stability of the binding pose generated from IFD for the identified hit inhibitor 6 and to investigate any potential interactions with conserved structural water molecules within the Pim-1 ATP binding site, a 500 ns MD simulation was conducted. CHARMM force field parameters for ligands were generated using the ParamChem CGenFF interface and refined based on MP2/6-31G* identified global minima. As shown in Figure S2 (Supplementary), parametrization of compound 6 force field generated by ParamChem CGenFF interface resulted in a close superposition between the CHARMM-minimized structure (purple) and the MP2-derived global minima (green), indicating high-quality force field fitting.31 Throughout the simulation, compound 6 formed stable interactions with key Pim-1 active site residues, maintaining its resulting IFD pose within the ATP binding pocket (Figure 4A and B). H-bond analysis revealed persistent interactions between the compound and K67 and E121, together with interactions involving structural water molecules located within the catalytic site and buried region. Interestingly, the pyrazole ring of compound 6, situated at the catalytic site, forms dual polar interactions with K67 and E89, functioning as both an H-bond donor and acceptor. This dual interaction is an uncommon feature among most reported Pim-1 inhibitors, which typically incorporate only an H-bond acceptor within their catalytic site-interacting moiety (Table 1). In comparison, inhibitors possessing both donor and acceptor functionalities in their catalytic site-interacting moiety, such as the oxindole-based inhibitor CX-6258, has their H-bonding group interacting indirectly with the backbone of D186 (PDB ID: 5O13).37 A 500 ns molecular dynamics simulation of compound 6 at Pim-1 catalytic site. (A) Binding pose of the active compound retrieved from the simulation and showing interacting water molecules. (B) H-bond analysis; dark blue: indolone nitrogen with E121 (hinge region), green: pyrazole nitrogen with K67, red: indolinone carbonyl with water, yellow: pyrazole-NH with structural water, sky blue: pyrazole NH with E89. B. The oxindole ring in compound 6, on the other hand, occupies the adenine binding region, forming a direct H-bond interaction with the backbone of hinge residue E121, as well as hydrophobic interactions across the hinge domain. Additionally, it is interacting with a structural water molecule located in the buried region near the gatekeeper residue L120, further enhancing the stability of the compound within the catalytic site. While the initial docking pose predicted aromatic stacking between compound 6 and F49 in the G-loop in its closed conformation, the MD simulation shows a transient opening of the G-loop. This conformational change, on the other hand, results in the formation of an H-bond between the G-loop backbone and the compound’s methoxy group, without compromising the overall stability of the ligand within the binding pocket (Figure 4A, 4B, S3 and Supplementary). In order to compare the stability of the protein in the presence and absence of the active compound, simulations were also performed for apo Pim-1 and Pim-1 in complex with compound 5, an inactive compound predicted to have high activity based on our docking results. RMSD analysis of backbone residues for both the whole protein and the active site region showed that compound 6 conferred higher structural stability of the active site compared to the apo form of Pim-1, with an average RMSD of 0.5 Å for active site residues, compared to 1.7 Å in the apo protein (Figure 5). Similarly, root-mean-square fluctuation (RMSF) analysis for active site residues over the full simulation period revealed lower RMSF values in the protein-compound 6 complex, particularly within the G-loop region (residues G45 to V52), and comparable stability in the hinge region (residues L120 to P125) relative to the apo protein. These findings suggest enhanced stabilization of the Pim-1 active site in the presence of the active compound 6 (Figure 6). RMSD for backbone residues for the (A) full protein and (B) active site residues from three different simulations comparing the apo protein, Pim-1 in complex with the inactive compound 5, and Pim-1 in complex with the active compound 6, throughout a 500 ns simulation for each system. Comparing RMSF for active site residues in three different simulations of the apo Pim-1, Pim1- in complex with the active compound 6, and Pim-1 in complex with the inactive compound 5 for amino acid residues in the range of 33 to 300. On the other hand, simulation of the Pim-1 in complex with the inactive compound 5 displayed a gradual increase in the RMSD for the whole protein throughout the 500 ns simulation period, with an average RMSD of 2.7 Å, and an average RMSD value for the active site of 1.5 Å. RMSF calculations for the active site residues show a reduced RMSF value in the G-loop region, yet it shows increased RMSF values for catalytic site residues and near the acidic patch domain, indicating weaker stabilizing interactions (Figures 5 and 6). Moreover, although pyrazole- and indolinone-containing chemotypes have individually been reported as Pim-1 inhibitors3,8,9 (Table 1), to the best of our knowledge, the pyrazolyl-methylene-indolinone scaffold represented by compound 6 has not previously been reported for Pim-1 inhibition. Compound 6 presents a distinct arrangement in which the pyrazole forms dual interaction with the catalytic region while the indolinone engages the hinge region. This scaffold may therefore provide a starting point for further structure-guided optimization toward improved Pim-1 inhibitory activity. Quantum Mechanics Calculations While MD simulations indicated greater active site stability in the presence of the active compound 6 compared to the inactive compound 5, further investigations were done using quantum mechanical calculations to better understand the in vitro findings obtained. Notably, compound 6 was the only compound to exhibit measurable Pim-1 inhibition, despite the docking predictions favoring the thienopyrimidine-containing compound 5, which showed the highest docking scores and favorable ionic interactions with the acidic patch when compared to compound 6 (Figure 7A, B and Table 2). (A) Docking pose of the highest docking score (compound 5). (B) Docking pose of active compound 6. (C) Co-planarity between the catalytic region-binding domain (black-dotted circle) and the aromatic ring that binds near the hinge region (black-color circle) is prevalent as revealed through the superposition of a set of pim-1 co-crystallized inhibitors. Since the polar interaction with catalytic K67 is a key determinant and an important feature in Pim-1 molecular recognition, and as the experimental pKa value for the thienopyrimidine ring has not been reported before, the electrostatic potential maps for compounds 5 and 6 were calculated. This would enable a comparison of the partial charges of the K67-binding groups in compounds 5 and 6. The HF/6-31G* calculations used to find the partial charges in water revealed higher electronic density on the thienopyrimidine nitrogen in compound 5 (−0.749) than on the pyrazole nitrogen in compound 6 (−0.576) (Figure S4 and Supplementary), suggesting both compounds possess sufficient charge density to engage in favorable polar interactions with catalytic K67, assuming favorable distances and orientations are achieved. To further probe the energetic basis of ligand binding, DFT-based interaction energy calculations were performed. Selected frames representing Protein-compound complexes were retrieved from the MD simulation runs for the active and the inactive compounds (6 and 5, respectively), where they were first fully optimized using the ONIOM model.2 The ligands were treated at the M062X/6-311+G(d)3 level of theory, while the protein residues were treated at the M062X/6-31G(d)4 level. The fully optimized structures were then used to compute basis set superposition error (BSSE)5 and the corrected interaction energies between the ligand and the protein at the M062X/6-311+G(d) level of theory. The resulting BSSE-corrected interaction energies were −78.8 kcal/mol for the active compound 6 model and −202.9 kcal/mol for the inactive compound 5 model. Such results suggest that the interaction energies of the inactive compound are higher compared to the active compound and, similar to the docking scores generated from force field molecular mechanics, could not justify the activities obtained of tested compounds. Conformational Prerequisite: Co-Planarity Between the Catalytic Lysine-Binding Moiety and the Hinge-Region-Binding Domain is Required for Effective Pim-1 Inhibitory Activity To further understand the discrepancy between the favorable docking scores and the experimental inhibition data, a conformational analysis of the active and inactive compounds was conducted. A prevalent structural feature observed among co-crystallized Pim-1 inhibitors is the co-planar arrangement between the catalytic-region-binding moiety and the hinge-region-binding moiety (Figure 7C). Accordingly, QM dihedral energy scans were performed for compounds 5 and 6 analyze their conformational preferences and their energetic requirements for coplanar geometry. In the gas phase, MP2/6-31G* dihedral scans for compound 5 revealed an optimal dihedral angle of approximately 50° between the thienopyrimidine ring (K67-interacting moiety) and the phenyl ring (hinge-region occupying moiety). However, adoption of a coplanar conformation was energetically unfavorable due to steric interactions involving the piperidine substituent, resulting in an energy penalty of approximately 8 kcal/mol. Notably, the IFD-derived pose of compound 5 did not adopt this high-energy coplanar geometry. Instead, compound 5 retained a non-planar conformation, with a dihedral angle of approximately 35°, corresponding to a relatively low-energy state located only ~2 kcal/mol above the calculated global minimum (Figure 8A). This energetically accessible conformation, together with the favorable electrostatic interactions established with the acidic patch, may account for the favorable docking scores obtained for this experimentally inactive compound. Gas phase MP2/6-31G* dihedral energy scans for compounds 5 and 6. The targeted dihedral is marked with a red curved arrow. Colored asterisks on the graphs correlate to the colored conformation. (A) Docking conformation of 5 (Orange) compared to a coplanar geometry conformation between the two aromatic rings (blue). (B) Docking conformation of 6, which represents the lowest energy conformation and a coplanar conformation between the two aromatic rings. In contrast, compound 6 adopts the coplanar geometry in its lowest-energy state, with the pyrazole (K67-interacting moiety) and the oxindole (hinge-region interacting moiety) rings co-planar. This geometry was preserved in its IFD docking pose (Figure 8B), suggesting an energetically accessible conformation conducive to binding. Although the non-coplanar conformation of compound 5 could be accommodated within the Pim-1 binding site without apparent steric clashes, the energy cost of achieving a coplanar geometry may hinder ligand access to the catalytic site, potentially limiting its inhibitory capacity. Compound 5 would therefore incur a substantial energetic penalty to approach such a geometry, whereas compound 6 intrinsically favors a near-coplanar arrangement. Conventional docking does not explicitly capture this potential conformational requirement, as the algorithm evaluates favorable final binding poses without explicitly accounting for the conformational accessibility of the ligand during entry into the binding pocket. Collectively, the results suggest that the energetic accessibility of near-coplanar conformations between the catalytic-region- and hinge-binding moieties may represent an important structural prerequisite for productive Pim-1 binding and inhibition. Conclusion Computational techniques in drug discovery rely heavily on the use of molecular mechanics force fields that, despite tremendous advances in describing a wide range of chemical groups present in drug-like molecules, their output should be carefully filtered and re-evaluated based on the specific molecular target and its binding requirements. In the absence structural information for protein-ligand complexes, the use of QM calculations can be advantageous in the field of drug discovery. This advantage is due to the capability of QM calculations to predict the global minima and to estimate the energy expense of conformational changes during the ligand binding. In the pursuit of a new chemical scaffold with Pim-1 inhibitory activity, a virtual screening approach was employed to retrieve a set of potential inhibitors. Subsequently, these compounds were subjected to in vitro evaluation that served as a basis for conducting MD simulations and QM calculations in order to investigate the structural determinants for inhibiting Pim-1 activity. The present study highlights the energetic accessibility of co-planarity between the moiety interacting with the catalytic lysine and the aromatic system occupying the hinge region as a potential structural prerequisite for productive Pim-1 inhibition. This work led to the identification of a 1H-pyrazolyl-methylene-indolinone derivative (compound 6) as a Pim-1 inhibitory hit. MD simulations revealed stable and distinct interactions between this compound and the Pim-1 active site, providing mechanistic insight into its inhibitory activity. The aforementioned findings have enhanced our comprehension of the structural prerequisites for effective Pim-1 inhibition, thereby facilitating the development of new Pim-1 inhibitors. The identified pyrazolyl-methylene-indolinone hit provides a previously unreported scaffold for Pim-1 inhibition that may serve as a starting point for future optimization toward improved inhibitory activity. Acknowledgments The authors would like to express their sincere gratitude and appreciation for the opportunity they have had to use the UNC research-computing cluster, The University of North Carolina at Chapel Hill. Funding Statement This research was funded by the Deanship of Scientific Research at Jordan University of Science and Technology, grant numbers 20200630 and 20220244. Author Contributions All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work. Disclosure The authors report there are no conflicts of interest in this work. References