Estimation of Drug-Target Residence Time by Targeted Molecular Dynamics Simulations

The importance of binding kinetics and drug-target residence time in pharmacology

A dominant assumption in pharmacology throughout the 20th century has been that in vivo target occupancy-and attendant pharmacodynamics-depends on the systemic concentration of drug relative to the equilibrium dissociation constant for the drug-target complex. In turn, the duration of pharmacodynamics is temporally linked to the systemic pharmacokinetics of the drug. Yet, there are many examples of drugs for which pharmacodynamic effect endures long after the systemic concentration of a drug has waned to (equilibrium) insignificant levels. To reconcile such data, the drug-target residence time model was formulated, positing that it is the lifetime (or residence time) of the binary drug-target complex, and not its equilibrium affinity per se, that determines the extent and duration of drug pharmacodynamics. Here, we review this model, its evolution over time, and its applications to natural ligand-macromolecule biology and synthetic drug-target pharmacology.

Alanine Scanning to Define Membrane Protein-Lipid Interaction Sites Using Native Mass Spectrometry

Lipids surrounding membrane proteins interact with different sites on the protein at varying specificities, ranging from highly specific to weak interactions. These interactions can modulate the structure, function, and stability of membrane proteins. Thus, to better understand membrane protein structure and function, it is important to identify the locations of lipid binding and the relative specificities of lipid binding at these sites. In our previous native mass spectrometry (MS) study, we developed a single and double mutant analysis approach to profile the contribution of specific residues toward lipid binding. Here, we extend this method by screening a broad range of mutants of AqpZ to identify specific lipid binding sites and by measuring binding of different lipid types to measure the selectivity of different lipids at selected binding sites. We complemented these native MS studies with molecular dynamics (MD) simulations to visualize lipid interactions at selected sites. We discovered that AqpZ is selective towards cardiolipins (CL) but only at specific sites. Specifically, CL orients with its headgroup facing the cytoplasmic side, and its acyl chains interact with a hydrophobic pocket located at the monomeric interface within the lipid bilayer. Overall, this integrative approach provides unique insights into lipid binding sites and the selectivity of various lipids towards AqpZ, enabling us to map the AqpZ protein structure based on the lipid affinity.

Development of novel CDK9 and CYP3A4 inhibitors for cancer therapy through field and computational approaches

Cyclin-dependent kinase 9 (CDK9) and cytochrome P450 3A4 (CYP3A4) have emerged as promising targets in the development of anticancer drugs, presenting a consistent challenge in the quest for potent inhibitors. CDK9 inhibitors can selectively target fast-growing cancer cells by disrupting transcription elongation, which in turn hinders the production of proteins essential for cell cycle progression and survivaŚ. Understanding how CYP3A4 metabolizes specific chemotherapy drugs allows for personalized treatment plans, optimizing drug dosages according to a patient's metabolic profile. Since many cancer patients undergo combination therapies, and CYP3A4 is vital in drug metabolism, its inhibition or induction by one drug can alter the plasma levels of others, potentially leading to treatment failure or increased toxicity. Therefore, managing CYP3A4 activity is critical for effective cancer treatment. Employing a range of computational methodologies, this study systematically investigated the binding mechanisms of pyrimidine derivatives against CDK9 and CYP3A4. The field-based model demonstrated high R 2 values (0.99), with Q2 (0.66), demonstrating its ability to predict in silico inhibitory activity against the target of this study. The screening process followed in this work led to the discovery of powerful new inhibitor compounds. Of the 15 new compounds designed, three have a high affinity with the target (ranging from -8 to -9 kcal/mol kcal/mol) and were singled out through docking filtration for more detailed investigation. As well as, a reference compound with a substantial pIC50 value of 8.4, serving as the foundation for the development of the new compounds, was included for comparative analysis. To elucidate the essential features of CDK9 and CYP3A4 inhibitor design, a comparative analysis was conducted between 3D-QSAR-generated contours and molecular docking conformations of ligands. Molecular dynamics simulations were carried out for a duration of 100 ns on selected docked complexes, specifically those involving novel compounds with CDK9 and CYP3A4 enzymes. Additionally, the binding free energy for these complexes was assessed using the MM/PBSA method, which evaluates the free energy landscape of protein-ligand interactions. The results of MM/PBSA highlighted the strength of the new compounds in enhancing interactions with the target protein, which favors the results of molecular docking and MD simulation. These insights contribute to a deeper understanding of the mechanisms underlying CDK9 and CYP3A4 inhibition, offering potential avenues for the development of innovative and effective CDK9 inhibitors.

Increased throughput in methods for simulating protein ligand binding and unbinding

By incorporating full flexibility and enabling the quantification of crucial parameters such as binding free energies and residence times, methods for investigating protein-ligand binding and unbinding via molecular dynamics provide details on the involved mechanisms at the molecular level. While these advancements hold promise for impacting drug discovery, a notable drawback persists: their relatively time-consuming nature limits throughput. Herein, we survey recent implementations which, employing a blend of enhanced sampling techniques, a clever choice of collective variables, and often machine learning, strive to enhance the efficiency of new and previously reported methods without compromising accuracy. Particularly noteworthy is the validation of these methods that was often performed on systems mirroring real-world drug discovery scenarios.

3D computer modeling of inhibitors targeting the MCF-7 breast cancer cell line

This study focused on developing new inhibitors for the MCF-7 cell line to contribute to our understanding of breast cancer biology and various experimental techniques. 3D QSAR modeling was used to design new tetrahydrobenzo[4, 5]thieno[2, 3-d]pyrimidine derivatives with good characteristics. Two robust 3D-QSAR models were developed, and their predictive capacities were confirmed through high correlations [CoMFA (Q2 = 0.62, R 2 = 0.90) and CoMSIA (Q2 = 0.71, R 2 = 0.88)] via external validations (R2 ext = 0.90 and R2 ext = 0.91, respectively). These successful evaluations confirm the potential of the models to provide reliable predictions. Six candidate inhibitors were discovered, and two new inhibitors were developed in silico using computational methods. The ADME-Tox properties and pharmacokinetic characteristics of the new derivatives were evaluated carefully. The interactions between the new tetrahydrobenzo[4, 5]thieno[2, 3-d]pyrimidine derivatives and the protein ERα (PDB code: 4XO6) were highlighted by molecular docking. Additionally, MM/GBSA calculations and molecular dynamics simulations provided interesting information on the binding stabilities between the complexes. The pharmaceutical characteristics, interactions with protein, and stabilities of the inhibitors were examined using various methods, including molecular docking and molecular dynamics simulations over 100 ns, binding free energy calculations, and ADME-Tox predictions, and compared with the FDA-approved drug capivasertib. The findings indicate that the inhibitors exhibit significant binding affinities, robust stabilities, and desirable pharmaceutical characteristics. These newly developed compounds, which act as inhibitors to mitigate breast cancer, therefore possess considerable potential as prospective drug candidates.

Computational 3D Modeling-Based Identification of Inhibitors Targeting Cysteine Covalent Bond Catalysts for JAK3 and CYP3A4 Enzymes in the Treatment of Rheumatoid Arthritis

This work aimed to find new inhibitors of the CYP3A4 and JAK3 enzymes, which are significant players in autoimmune diseases such as rheumatoid arthritis. Advanced computer-aided drug design techniques, such as pharmacophore and 3D-QSAR modeling, were used. Two strong 3D-QSAR models were created, and their predictive power was validated by the strong correlation (R2 values > 80%) between the predicted and experimental activity. With an ROC value of 0.9, a pharmacophore model grounded in the DHRRR hypothesis likewise demonstrated strong predictive ability. Eight possible inhibitors were found, and six new inhibitors were designed in silico using these computational models. The pharmacokinetic and safety characteristics of these candidates were thoroughly assessed. The possible interactions between the inhibitors and the target enzymes were made clear via molecular docking. Furthermore, MM/GBSA computations and molecular dynamics simulations offered insightful information about the stability of the binding between inhibitors and CYP3A4 or JAK3. Through the integration of various computational approaches, this study successfully identified potential inhibitor candidates for additional investigation and efficiently screened compounds. The findings contribute to our knowledge of enzyme-inhibitor interactions and may help us create more effective treatments for autoimmune conditions like rheumatoid arthritis.

The Impact of Data on Structure-Based Binding Affinity Predictions Using Deep Neural Networks

Artificial intelligence (AI) has gained significant traction in the field of drug discovery, with deep learning (DL) algorithms playing a crucial role in predicting protein-ligand binding affinities. Despite advancements in neural network architectures, system representation, and training techniques, the performance of DL affinity prediction has reached a plateau, prompting the question of whether it is truly solved or if the current performance is overly optimistic and reliant on biased, easily predictable data. Like other DL-related problems, this issue seems to stem from the training and test sets used when building the models. In this work, we investigate the impact of several parameters related to the input data on the performance of neural network affinity prediction models. Notably, we identify the size of the binding pocket as a critical factor influencing the performance of our statistical models; furthermore, it is more important to train a model with as much data as possible than to restrict the training to only high-quality datasets. Finally, we also confirm the bias in the typically used current test sets. Therefore, several types of evaluation and benchmarking are required to understand models' decision-making processes and accurately compare the performance of models.

Predicting Protein-Ligand Binding and Unbinding Kinetics with Biased MD Simulations and Coarse-Graining of Dynamics: Current State and Challenges

The prediction of drug-target binding and unbinding kinetics that occur on time scales between milliseconds and several hours is a prime challenge for biased molecular dynamics simulation approaches. This Perspective gives a concise summary of the theory and the current state-of-the-art of such predictions via biased simulations, of insights into the molecular mechanisms defining binding and unbinding kinetics as well as of the extraordinary challenges predictions of ligand kinetics pose in comparison to binding free energy predictions.