Protein structures in virtual screening: a case study with CDK2

Design, synthesis and evaluation of thieno[3,2-d]pyrimidine derivatives as novel potent CDK7 inhibitors

The targeting of cyclin-dependent kinase 7 (CDK7) has become a highly desirable therapeutic approach in the field of oncology due to its dual role in regulating essential biological processes, encompassing cell cycle progression and transcriptional control. We have previously identified a highly selective thieno[3,2-d]pyrimidine-based CDK7 inhibitor with demonstrated efficacy and safety in animal model. In this study, we sought to optimize the thieno[3,2-d]pyrimidine core to discover a novel series of CDK7 inhibitors with improved potency and pharmacokinetic (PK) properties. Through extensive structure-activity relationship (SAR) studies, compound 20 has emerged as the lead candidate due to its potent inhibitory activity against CDK7 and remarkable efficacy on MDA-MB-453 cells, a representative triple negative breast cancer (TNBC) cell line. Furthermore, 20 has demonstrated favorable oral bioavailability and exhibited highly desirable pharmacokinetic (PK) properties, making it a promising lead candidate for further structural optimization.

Discovery and optimization of thieno[3,2-d]pyrimidine derivatives as highly selective inhibitors of cyclin-dependent kinase 7

Targeting cyclin-dependent kinase 7 (CDK7) has emerged as a highly sought-after therapeutic strategy in oncology due to its duality of function in regulating biological processes, including cell cycle progression and transcriptional control. Herein, we describe the design, optimization and characterization of a series of thieno[3,2-d]pyrimidine derivatives as potent CDK7 inhibitors. The involvement of thiophene as core structure plays critical role in leading to the remarkable selectivity and incorporation of a fluorine atom into the piperidine ring enhances metabolic stability. Structure-activity relationship (SAR) study generated compound 36 as lead compound with potent inhibitory activity against CDK7 and good kinome selectivity in vitro. Compound 36 demonstrated strong efficacy against a triple negative breast cancer (TNBC) cell line-derived xenograft (CDX) mouse model upon oral administration at 5 mg/kg once daily. Therefore, it exhibits immense potential as a lead candidate for further exploration in the development of cancer therapy.

Improving Structure-Based Virtual Screening with Ensemble Docking and Machine Learning

One of the main challenges of structure-based virtual screening (SBVS) is the incorporation of the receptor's flexibility, as its explicit representation in every docking run implies a high computational cost. Therefore, a common alternative to include the receptor's flexibility is the approach known as ensemble docking. Ensemble docking consists of using a set of receptor conformations and performing the docking assays over each of them. However, there is still no agreement on how to combine the ensemble docking results to obtain the final ligand ranking. A common choice is to use consensus strategies to aggregate the ensemble docking scores, but these strategies exhibit slight improvement regarding the single-structure approach. Here, we claim that using machine learning (ML) methodologies over the ensemble docking results could improve the predictive power of SBVS. To test this hypothesis, four proteins were selected as study cases: CDK2, FXa, EGFR, and HSP90. Protein conformational ensembles were built from crystallographic structures, whereas the evaluated compound library comprised up to three benchmarking data sets (DUD, DEKOIS 2.0, and CSAR-2012) and cocrystallized molecules. Ensemble docking results were processed through 30 repetitions of 4-fold cross-validation to train and validate two ML classifiers: logistic regression and gradient boosting trees. Our results indicate that the ML classifiers significantly outperform traditional consensus strategies and even the best performance case achieved with single-structure docking. We provide statistical evidence that supports the effectiveness of ML to improve the ensemble docking performance.

Inhibition of Oncogenic Kinases: An In Vitro Validated Computational Approach Identified Potential Multi-Target Anticancer Compounds

Tumorigenesis in humans is a multistep progression that imitates genetic changes leading to cell transformation and malignancy. Oncogenic kinases play a central role in cancer progression, rendering them putative targets for the design of anti-cancer drugs. The presented work aims to identify the potential multi-target inhibitors of oncogenic receptor tyrosine kinases (RTKs) and serine/threonine kinases (STKs). For this, chemoinformatics and structure-based virtual screening approaches were combined with an in vitro validation of lead hits on both cancerous and non-cancerous cell lines. A total of 16 different kinase structures were screened against ~739,000 prefiltered compounds using diversity selection, after which the top hits were filtered for promising pharmacokinetic properties. This led to the identification of 12 and 9 compounds against RTKs and STKs, respectively. Molecular dynamics (MD) simulations were carried out to better comprehend the stability of the predicted hit kinase-compound complexes. Two top-ranked compounds against each kinase class were tested in vitro for cytotoxicity, with compound F34 showing the most promising inhibitory activity in HeLa, HepG2, and Vero cell lines with IC50 values of 145.46 μM, 175.48 μM, and 130.52 μM, respectively. Additional docking of F34 against various RTKs was carried out to support potential multi-target inhibition. Together with reliable MD simulations, these results suggest the promising potential of identified multi-target STK and RTK scaffolds for further kinase-specific anti-cancer drug development toward combinatorial therapies.

Online structure-based screening of purchasable approved drugs and natural compounds: retrospective examples of drug repositioning on cancer targets

Drug discovery is a long and difficult process that benefits from the integration of virtual screening methods in experimental screening campaigns such as to generate testable hypotheses, accelerate and/or reduce the cost of drug development. Current drug attrition rate is still a major issue in all therapeutic areas and especially in the field of cancer. Drug repositioning as well as the screening of natural compounds constitute promising approaches to accelerate and improve the success rate of drug discovery. We developed three compounds libraries of purchasable compounds: Drugs-lib, FOOD-lib and NP-lib that contain approved drugs, food constituents and natural products, respectively, that are optimized for structure-based virtual screening studies. The three compounds libraries are implemented in the MTiOpenScreen web server that allows users to perform structure-based virtual screening computations on their selected protein targets. The server outputs a list of 1,500 molecules with predicted binding scores that can then be processed further by the users and purchased for experimental validation. To illustrate the potential of our service for drug repositioning endeavours, we selected five recently published drugs that have been repositioned in vitro and/or in vivo on cancer targets. For each drug, we used the MTiOpenScreen service to screen the Drugs-lib collection against the corresponding anti-cancer target and we show that our protocol is able to rank these drugs within the top ranked compounds. This web server should assist the discovery of promising molecules that could benefit patients, with faster development times, and reduced costs and risk.

Recent advances on CDK inhibitors: An insight by means of in silico methods

The cyclin dependent kinases (CDKs) are a small family of serine/threonine protein kinases that can act as a potential therapeutic target in several proliferative diseases, including cancer. This short review is a survey on the more recent research progresses in the field achieved by using in silico methods. All the "armamentarium" available to the medicinal chemists (docking protocols and molecular dynamics, fragment-based, de novo design, virtual screening, and QSAR) has been employed to the discovery of new, potent, and selective inhibitors of cyclin dependent kinases. The results cited herein can be useful to understand the nature of the inhibitor-target interactions, and furnish an insight on the structural/molecular requirements necessary to achieve the required selectivity against cyclin dependent kinases over other types of kinases.

Visualizing ensembles in structural biology

Displaying a single representative conformation of a biopolymer rather than an ensemble of states mistakenly conveys a static nature rather than the actual dynamic personality of biopolymers. However, there are few apparent options due to the fixed nature of print media. Here we suggest a standardized methodology for visually indicating the distribution width, standard deviation and uncertainty of ensembles of states with little loss of the visual simplicity of displaying a single representative conformation. Of particular note is that the visualization method employed clearly distinguishes between isotropic and anisotropic motion of polymer subunits. We also apply this method to ligand binding, suggesting a way to indicate the expected error in many high throughput docking programs when visualizing the structural spread of the output. We provide several examples in the context of nucleic acids and proteins with particular insights gained via this method. Such examples include investigating a therapeutic polymer of FdUMP (5-fluoro-2-deoxyuridine-5-O-monophosphate) - a topoisomerase-1 (Top1), apoptosis-inducing poison - and nucleotide-binding proteins responsible for ATP hydrolysis from Bacillus subtilis. We also discuss how these methods can be extended to any macromolecular data set with an underlying distribution, including experimental data such as NMR structures.

Benchmarking Data Sets for the Evaluation of Virtual Ligand Screening Methods: Review and Perspectives

Virtual screening methods are commonly used nowadays in drug discovery processes. However, to ensure their reliability, they have to be carefully evaluated. The evaluation of these methods is often realized in a retrospective way, notably by studying the enrichment of benchmarking data sets. To this purpose, numerous benchmarking data sets were developed over the years, and the resulting improvements led to the availability of high quality benchmarking data sets. However, some points still have to be considered in the selection of the active compounds, decoys, and protein structures to obtain optimal benchmarking data sets.

Multiple structures for virtual ligand screening: defining binding site properties-based criteria to optimize the selection of the query

Structure based virtual ligand screening (SBVLS) methods are widely used in drug discovery programs. When several structures of the target are available, protocols based either on single structure docking or on ensemble docking can be used. The performance of the methods depends on the structure(s) used as a reference, whose choice requires retrospective enrichment studies on benchmarking databases which consume additional resources. In the present study, we have identified several trends in the properties of the binding sites of the structures that led to the optimal performance in retrospective SBVLS tests whatever the docking program used (Surflex-dock or ICM). By assessing their hydrophobicity and comparing their volume and opening, we show that the selection of optimal structures should be possible with no requirement of prior retrospective enrichment studies. If the mean binding site volume is lower than 350 A(3), the structure with the smaller volume should be preferred. In the other cases, the structure with the largest binding site should be preferred. These optimal structures may be either selected for a single structure docking strategy or an ensemble docking strategy. When constructing an ensemble, the opening of the site might be an interesting criterion additionaly to its volume as the most closed structures should not be preferred in the large systems. These "binding site properties-based" guidelines could be helpful to optimize future prospective drug discovery protocols when several structures of the target are available.

Discovery of novel checkpoint kinase 1 inhibitors by virtual screening based on multiple crystal structures

Incorporating receptor flexibility is considered crucial for improvement of docking-based virtual screening. With an abundance of crystallographic structures freely available, docking with multiple crystal structures is believed to be a practical approach to cope with protein flexibility. Here we describe a successful application of the docking of multiple structures to discover novel and potent Chk1 inhibitors. Forty-six Chk1 structures were first compared in single structure docking by predicting the binding mode and recovering known ligands. Combinations of different protein structures were then compared by recovery of known ligands and an optimal ensemble of Chk1 structures were selected. The chosen structures were used in the virtual screening of over 60 000 diverse compounds for Chk1 inhibitors. Six novel compounds ranked at the top of the hits list were tested experimentally, and two of these compounds inhibited Chk1 activity-the best with an IC(50) value of 9.6 μM. Further study indicated that achieving a better enrichment and identifying more diverse compounds was more likely using multiple structures than using only a single structure even when protein structures were randomly selected. Taking into account conformational energy difference did not help to improve enrichment in the top ranked list.

Molecular recognition in the case of flexible targets

A protein's flexibility is well recognized to underlie its capacity to engage in critical functions, such as signal transduction, biomolecular transport and biochemical reactivity. Molecular recognition is also tightly linked to the dynamics of the binding partners, yet protein flexibility has largely been ignored by the growing field of structure-based drug design (SBDD). In combination with experimentally determined structures, a number of computational methods have been proposed to model protein movements, which may be important for small molecule binding. Such techniques have the ability to expose new binding site conformations, which may in turn recognize and lead to the discovery of more potent and selective drugs through molecular docking. In this article, we discuss various methods and focus on the Relaxed Complex Scheme (RCS), which uses Molecular Dynamics (MD) simulations to model full protein flexibility and enhance virtual screening programmes. We review practical applications of the RCS and use a recent study of the HIV-1 reverse transcriptase to illustrate the various phases of the scheme. We also discuss some encouraging developments, aimed at addressing current weaknesses of the RCS.

Molecular mechanism of serotonin transporter inhibition elucidated by a new flexible docking protocol

The two main groups of antidepressant drugs, the tricyclic antidepressants (TCAs) and the selective serotonin reuptake inhibitors (SSRIs), as well as several other compounds, act by inhibiting the serotonin transporter (SERT). However, the binding mode and molecular mechanism of inhibition in SERT are not fully understood. In this study, five classes of SERT inhibitors were docked into an outward-facing SERT homology model using a new 4D ensemble docking protocol. Unlike other docking protocols, where protein flexibility is not considered or is highly dependent on the ligand structure, flexibility was here obtained by side chain sampling of the amino acids of the binding pocket using biased probability Monte Carlo (BPMC) prior to docking. This resulted in the generation of multiple binding pocket conformations that the ligands were docked into. The docking results showed that the inhibitors were stacked between the aromatic amino acids of the extracellular gate (Y176, F335) presumably preventing its closure. The inhibitors interacted with amino acids in both the putative substrate binding site and more extracellular regions of the protein. A general structure-docking-based pharmacophore model was generated to explain binding of all studied classes of SERT inhibitors. Docking of a test set of actives and decoys furthermore showed that the outward-facing ensemble SERT homology model consistently and selectively scored the majority of active compounds above decoys, which indicates its usefulness in virtual screening.

[Molecular docking: role of intermolecular contacts in formation of complexes of proteins with nucleotides and peptides]

Knowledge of 3D-structure of protein-ligand complex is a major prerequisite for understanding the functioning mechanism of cellular proteins and membrane receptors. This is also of a great help in rational drug design projects. In the present paper we briefly review the molecular docking approaches used to predict possible orientation of a ligand in the protein binding site. The recent trends to improve the accuracy and efficiency of docking algorithms are demonstrated with the results obtained in Laboratory of Biomolecular Modeling. Particular attention is paid to protein-ligand hydrophobic and stacking interactions responsible for molecular recognition of ligand fragments. Such type of interactions are not always adequately represented in scoring criteria of docking applications that leads to mismatch in 3D-structure complexes predictions. That is why further inquiry of methods to account for these interactions is now the area of active research.

Recipes for the selection of experimental protein conformations for virtual screening

The use of multiple X-ray protein structures has been reported to be an efficient alternative for the representation of the binding pocket flexibility needed for accurate small molecules docking. However, the docking performance of the individual single conformations varies widely, and adding certain conformations to an ensemble is even counterproductive. Here we used a very large and diverse benchmark of 1068 X-ray protein conformations of 99 therapeutically relevant proteins, first, to compare the performance of the ensemble and single-conformation docking and, second, to find the properties of the best-performing conformers that can be used to select a smaller set of conformers for ensemble docking. The conformer selection has been validated through retrospective virtual screening experiments aimed at separating known ligand binders from decoys. We found that the conformers cocrystallized with the largest ligands displayed high selectivity for binders, and when combined in ensembles they consistently provided better results than randomly chosen protein conformations. The use of ensembles encompassing between 3 and 5 experimental conformations consistently improved the docking accuracy and binders vs decoys separation.

High-performance drug discovery: computational screening by combining docking and molecular dynamics simulations

Virtual compound screening using molecular docking is widely used in the discovery of new lead compounds for drug design. However, this method is not completely reliable and therefore unsatisfactory. In this study, we used massive molecular dynamics simulations of protein-ligand conformations obtained by molecular docking in order to improve the enrichment performance of molecular docking. Our screening approach employed the molecular mechanics/Poisson-Boltzmann and surface area method to estimate the binding free energies. For the top-ranking 1,000 compounds obtained by docking to a target protein, approximately 6,000 molecular dynamics simulations were performed using multiple docking poses in about a week. As a result, the enrichment performance of the top 100 compounds by our approach was improved by 1.6–4.0 times that of the enrichment performance of molecular dockings. This result indicates that the application of molecular dynamics simulations to virtual screening for lead discovery is both effective and practical. However, further optimization of the computational protocols is required for screening various target proteins.

Lead discovery is one of the most important processes in rational drug design. To improve the rate of the detection of lead compounds, various technologies such as high-throughput screening and combinatorial chemistry have been introduced into the pharmaceutical industry. However, since these technologies alone may not improve lead productivity, computational screening has become important. A central method for computational screening is molecular docking. This method generally docks many flexible ligands to a rigid protein and predicts the binding affinity for each ligand in a practical time. However, its ability to detect lead compounds is less reliable. In contrast, molecular dynamics simulations can treat both proteins and ligands in a flexible manner, directly estimate the effect of explicit water molecules, and provide more accurate binding affinity, although their computational costs and times are significantly greater than those of molecular docking. Therefore, we developed a special purpose computer “MDGRAPE-3” for molecular dynamics simulations and applied it to computational screening. In this paper, we report an effective method for computational screening; this method is a combination of molecular docking and massive-scale molecular dynamics simulations. The proposed method showed a higher and more stable enrichment performance than the molecular docking method used alone.

Consistent improvement of cross-docking results using binding site ensembles generated with elastic network normal modes

The representation of protein flexibility is still a challenge for the state-of-the-art flexible ligand docking protocols. In this article, we use a large and diverse benchmark to prove that is possible to improve consistently the cross-docking performance against a single receptor conformation, using an equilibrium ensemble of binding site conformers. The benchmark contained 28 proteins, and our method predicted the top-ranked near native ligand poses 20% more efficiently than using a single receptor. The multiple conformations were derived from the collective variable space defined by all heavy-atom elastic network normal modes, including backbone and side chains. We have found that the binding site displacements for best positioning of the ligand seem rather independent from the global collective motions of the protein. We also found that the number of binding site conformations needed to represent nonredundant flexibility was < 100. The ensemble of receptor conformations can be generated at our Web site at http://abagyan.scripps.edu/MRC.

Combining docking with pharmacophore filtering for improved virtual screening

BACKGROUND

Virtual screening is used to distinguish potential leads from inactive compounds in a database of chemical samples. One method for accomplishing this is by docking compounds into the structure of a receptor binding site in order to rank-order compounds by the quality of the interactions they form with the receptor. It is generally established that docking can be reasonably successful at generating good poses of a ligand in an active site. However, the scoring functions that are used with docking are typically not successful at correctly ranking ligands according to binding affinity or even distinguishing correct poses of a given ligand from incorrect ones.

RESULTS

We have developed a simple method for reducing the number of false positives in a virtual screen, meaning ligands which are scored highly by the docking program but do not bind well in reality. This method uses a docking program for pose generation without regard to scoring, followed by filtering with receptor-based pharmacophore searches. We applied it to three test-case targets: neuraminidase A, cyclin-dependent kinase 2, and the C1 domain of protein kinase C.

CONCLUSION

The pharmacophore filtering method can perform better than more traditional docking + scoring methods, and allows the advantages of both docking-based and pharmacophore-based approaches to virtual screening to be fully realized.

The identification of novel PLC-gamma inhibitors using virtual high throughput screening

Phospholipase C-gamma (PLC-gamma) has been identified as a possible biological target for anticancer drug therapy but suitable inhibitors are lacking. Therefore, in order to identify active compounds (hits) virtual high throughput screening was performed. The crystal structure of the PLC-delta isoform was used as a model docking scaffold since no crystallographic data are available on its gamma counterpart. A pilot screen was performed using approximately 9.2x10(4) compounds, where the robustness of the methodology was tested. This was followed by the main screening effort where approximately 4.4x10(5) compounds were used. In both cases, plausible compounds were identified (virtual hits) and a selection of these was experimentally tested. The most potent compounds were in the single digit micro-molar range as determined from the biochemical (Flashplate) assay. This translated into approximately 15 microM in a functional assay in cells. About 30% of the virtual hits showed activity against PLC-gamma (IC(50)<50 microM).

Synthesis and evaluation of pyrazolo[1,5-b]pyridazines as selective cyclin dependent kinase inhibitors

A novel series of pyrazolo[1,5-b]pyridazines have been synthesized and identified as cyclin dependant kinase inhibitors potentially useful for the treatment of solid tumors. Modification of the hinge-binding amine or the C(2)- and C(6)-substitutions on the pyrazolopyridazine core provided potent inhibitors of CDK4 and demonstrated enzyme selectivity against VEGFR-2 and GSK3beta.

Computational chemistry approaches to drug discovery in signal transduction

The advent of therapeutic strategies aimed at targeting specific macromolecular components of deregulated signaling pathways associated with particular disease states has given rise to the idea that it should be possible to design ligands as drug candidates to these targets from first principles. This concept has been beckoning for a long time but structure-based ligand design only became feasible once it was possible to determine the 3-D structures of molecular targets at atomic resolution. However, structure-based design turned out to be difficult, chiefly because under physiological conditions both receptors and ligands are not static but they behave dynamically. While it is possible to design ligands with high steric and electronic complementarity to a receptor site, it is always uncertain how biologically relevant the assumed conformations of both ligand and receptor actually are. The fact that it remains beyond our current abilities to predict with sufficient accuracy the affinity between hypothetical ligand and receptor poses is in part connected with this problem and continues to confound the reliable prediction of drug-like ligands for therapeutic targets. Nevertheless, significant progress has been made and so-called virtual screening methods that use computational methods to dock candidate ligands into receptor sites and to score the resulting complexes are now used routinely as one of the components in drug discovery screening campaigns. Here an overview is given of the underlying principles, implementations, and applications of structure-guided computational design technologies. Although the emphasis is on receptor-based strategies, mention will also be made of some of the more established ligand-based approaches, such as similarity analyses and quantitative structure-activity relationship methods.

Towards improving compound selection in structure-based virtual screening

Structure-based virtual screening is now an established technology for supporting hit finding and lead optimisation in drug discovery. Recent validation studies have highlighted the poor performance of currently used scoring functions in estimating binding affinity and hence in ranking large datasets of docked ligands. Progress in the analysis of large datasets can be made through the use of appropriate data mining techniques and the derivation of a broader range of descriptors relevant to receptor-ligand binding. In addition, simple scoring functions can be supplemented by simulation-based scoring protocols. Developments in workflow design allow the automation of repetitive tasks, and also encourage the routine use of simulation-based methods and the rapid prototyping of novel modelling and analysis procedures.

Discovery of a potent CDK2 inhibitor with a novel binding mode, using virtual screening and initial, structure-guided lead scoping

Virtual screening against a pCDK2/cyclin A crystal structure led to the identification of a potent and novel CDK2 inhibitor, which exhibited an unusual mode of interaction with the kinase binding motif. With the aid of X-ray crystallography and modelling, a medicinal chemistry strategy was implemented to probe the interactions seen in the crystal structure and to establish SAR. A fragment-based approach was also considered but a different, more conventional, binding mode was observed. Compound selectivity against GSK-3beta was improved using a rational design strategy, with crystallographic verification of the CDK2 binding mode.

A Multivariate Approach to Investigate Docking Parameters' Effects on Docking Performance

Increasingly powerful docking programs for analyzing and estimating the strength of protein-ligand interactions have been developed in recent decades, and they are now valuable tools in drug discovery. Software used to perform dockings relies on a number of parameters that affect various steps in the docking procedure. However, identifying the best choices of the settings for these parameters is often challenging. Therefore, the settings of the parameters are quite often left at their default values, even though scientists with long experience with a specific docking tool know that modifying certain parameters can improve the results. In the study presented here, we have used statistical experimental design and subsequent regression based on root-mean-square deviation values using partial least-square projections to latent structures (PLS) to scrutinize the effects of different parameters on the docking performance of two software packages: FRED and GOLD. Protein-ligand complexes with a high level of ligand diversity were selected from the PDBbind database for the study, using principal component analysis based on 1D and 2D descriptors, and space-filling design. The PLS models showed quantitative relationships between the docking parameters and the ability of the programs to reproduce the ligand crystallographic conformation. The PLS models also revealed which of the parameters and what parameter settings were important for the docking performance of the two programs. Furthermore, the variation in docking results obtained with specific parameter settings for different protein-ligand complexes in the diverse set examined indicates that there is great potential for optimizing the parameter settings for selected sets of proteins.

Evaluation of the utility of homology models in high throughput docking

High throughput docking (HTD) is routinely used for in silico screening of compound libraries with the aim to find novel leads in a drug discovery program. In the absence of an experimentally determined structure, a homology model can be used instead. Here we present an assessment of the utility of homology models in HTD by docking 300,000 anticipated inactive compounds along with 642 known actives into the binding site of the insulin-like growth factor 1 receptor (IGF-1R) kinase constructed by homology modeling. Twenty-one different templates were selected and the enrichment curves obtained by the homology models were compared to those obtained by three IGF-1R crystal structures. The results show a wide range of enrichments from random to as good as two of the three IGF-1R crystal structures. Nevertheless, if we consider the enrichment obtained at 2% of the database screened as a performance criterion, the best crystal structure outperforms the best homology model. Surprisingly, the sequence identity of the template to the target is not a good descriptor to predict the enrichment obtained by a homology model. The three homology models that yield the worst enrichment have the smallest binding-site volume. Based on our results, we propose ensemble docking to perform HTD with homology models.

Differential binding of inhibitors to active and inactive CDK2 provides insights for drug design

The cyclin-dependent kinases (CDKs) have been characterized in complex with a variety of inhibitors, but the majority of structures solved are in the inactive form. We have solved the structures of six inhibitors in both the monomeric CDK2 and binary CDK2/cyclinA complexes and demonstrate that significant differences in ligand binding occur depending on the activation state. The binding mode of two ligands in particular varies substantially in active and inactive CDK2. Furthermore, energetic analysis of CDK2/cyclin/inhibitors demonstrates that a good correlation exists between the in vitro potency and the calculated energies of interaction, but no such relationship exists for CDK2/inhibitor structures. These results confirm that monomeric CDK2 ligand complexes do not fully reflect active conformations, revealing significant implications for inhibitor design while also suggesting that the monomeric CDK2 conformation can be selectively inhibited.