Hidratáció a gyógyszertervezésben

Absztrakt

A structural and thermodynamic description of target-ligand interactions is a crucial aspect in drug design. In addition, water molecules in different hydration shells (bulk, surface, interface, buried) have various structural and energetic contributions in the complex. In this paper, we provide an overview of the calculation (prediction) of the structural and thermodynamic role of water molecules during ligand binding. Accurate determination of the spatial water positions in a target-ligand complex is not a trivial task. Experimental approaches often cannot provide void-free hydrated (complete) structures. Therefore, theoretical methods have been developed to complement experimental methods for calculating hydration problems in target-ligand complexes. Using the explicit water models is a good alternative to identify the role of each water molecule in ligand binding. For this purpose, molecular dynamics (MD)-based computational tools like MobyWat were introduced.

During computational docking of a ligand to the target, distinguishing between the conserved and replaced waters is also challenging. Protocols like HydroDock can answer these challenges.

Besides knowing the structure of the target-ligand complex, it is also important to calculate the target-ligand binding affinity (free energy of binding in terms of thermodynamics). In thermodynamic calculations, hydration also plays a key role.  In drug design, molecular docking of several thousand small molecules is commonly used to optimize the lead molecule. The rapid docking methods with implicit or without any water models lead to inaccurate results. Furthermore, especially in the case of larger, highly flexible, peptide-like ligands, the scoring functions of docking programs result in unrealistic binding affinities. If explicit water models are used in combination with MD calculations in which the contribution of the solvent itself to the target-ligand complex binding is also considered, more accurate thermodynamic quantities can be obtained. For example, in calculations performed with cyclodextrins, we separated the enthalpic contributions of three types of intermolecular interactions formed with the solvent (cyclodextrin-water, ligand-water, water-water), which are numerically comparable to the enthalpic contribution of the cyclodextrin-drug interaction. The entropic contribution associated with water is comparable in magnitude to the binding free energy, thus confirming the necessity of using an explicit water model for the accuracy of thermodynamics calculations. Using the results obtained, we also successfully studied the thermodynamic basis of the formation of cyclodextrin-based nanostructures.

As electronic contributions play a key role in the interaction between the partners, quantum mechanics (QM) can be considered the most precise approach from a theoretical point of view. Due to the continuous increase in computer performance, QM methods in drug design have recently gained more attention, especially in the calculation of target-ligand interactions. However, the large size of biomolecular systems still hinders the routine application of these methods. To overcome this, semi-empirical QM-based methods have been recently introduced for target-ligand calculations, mostly with PM7-like parametrization.

Although implicit water models (like COSMO) successfully handle the effects of the surrounding solvent medium in QM, target-ligand systems often contain specific water molecules that mediate interactions between partners, and explicit modeling of these molecules is essential. In our latest work, we performed PM7/COSMO calculations with explicit waters predicted by MobyWat. Our QMH-L method achieved good correlations in a wide size range for the ligands (up to an MW of 3300) with experimental thermodynamics values. Fast 1SCF calculations allow the use of QMH-L in a high-throughput manner.

The role of hydration has long been neglected in drug design, and most efforts have been made to optimize the interactions between the target molecule and the ligand. In recent decades, it has become obvious that this optimization process cannot function effectively without taking into account the role of individual water molecules. In this paper, we have given a quick overview of the role of hydration and the solution of practical problems arising in this regard during design, both at the structural and energetic levels. In addition to the widely used molecular mechanics methods, we have also discussed the most modern quantum mechanical approaches in calculating target-ligand binding affinities.