**1. Introduction**

To better understand the basis of the activity of any molecule with biological activity, it is important to know how this molecule interacts with its site of action, more specifically its conformational properties in solution and orientation for the interaction. Molecular recognition in biological systems relies on specific attractive and/or repulsive interactions between two partner molecules. This study seeks to identify such interactions between ligands and their host molecules, typically proteins, given their three-dimensional (3D) structures. Therefore, it is important to know about interaction geometries and approximate affinity contributions of attractive interactions. At the same time, it is necessary to be aware of the fact that molecular interactions behave in a highly non-additive fashion. The same interaction may account for different amounts of free energy in different contexts and any change in molecular structure might have multiple effects, so it is only reliable to compare similar structures. In fact, the multiple interactions present in a single two-molecule complex are a compromise between attractive and repulsive interactions. On the other hand, a molecular complex is not characterised by a single structure, as can be seen in crystal structures, but by an ensemble of structures. Furthermore, changes in the degree of freedom of both partners during an interaction have a large impact on binding free energy [Bissantz et al., 2010].

The availability of high-quality molecular graphics tools in the public domain is changing the way macromolecular structure is perceived by researchers, while computer modelling has emerged as a powerful tool for experimental and theoretical investigations. Visualisation of experimental data in a 3D, atomic-scale model can not only help to explain unexpected results but often raises new questions, thereby affecting future research. Models of sufficient quality can be set in motion in molecular dynamic (MD) simulations to move beyond a static picture and provide insight into the dynamics of important biological processes.

© 2012 Yunta, licensee InTech. This is an open access chapter distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. © 2012 Yunta, licensee InTech. This is a paper distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Computational methods have become increasingly important in a number of areas such as comparative or homology modelling, functional site location, characterisation of ligandbinding sites in proteins, docking of small molecules into protein binding sites, proteinprotein docking, and molecular dynamic simulations [see for example Choe & Chang, 2002]. Current results yield information that is sometimes beyond experimental possibilities and can be used to guide and improve a vast array of experiments.

Using Molecular Modelling to Study Interactions Between Molecules with Biological Activity 187

of substituents on a single bond may vary with time. Therefore, any compound containing one or several single bonds exists at every moment in many different conformers, but generally only low energy conformers are found to a large extent [Kund, 1997, as cited in

The biological activity of a drug molecule is supposed to depend on one single unique conformation amongst all the low energy conformations, the search for this so-called bioactive conformation for compound sets being one of the major tasks in Medicinal Chemistry. Searching for all low energy conformations is possible with molecular modelling studies, since molecular modelling is concerned with the description of the atomic and molecular interactions that govern microscopic and macroscopic behaviours of physical systems. These molecular interactions are classified as: (a) bonded (stretching, bending and torsion), (b) non-bonded (electrostatic (including interactions with metals), van der Waals and π-stacking), and (c) derived, as they result from the previous ones (hydrogen bonds and

Protein-ligand or, in general, molecule-molecule binding free energy differences can *a priori* be computed from first principles using free energy perturbation techniques and a full atomic detailed model with explicit solvent molecules using molecular dynamics simulations. However, these are computationally demanding. More affordable approaches use end-point molecular dynamic simulations and compute free energies accounting for solvent effects with continuum methods, such as MM-PBSA (molecular mechanics Poisson-Boltzman surface area) or MM-GBSA (generalized Born surface area) [Kollman et al., 2000; Wang et al., 2005]. One of the first approaches was comparative molecular field analysis (CoMFA) [Cramer et al., 1988], which enabled interpretation and understanding of enzyme active sites when the crystal structure was absent. However, this type of analysis was not possible until *in vitro* drug-drug interaction studies were widely used (through the 1990s).

Molecular mechanics (MM) is often the only feasible means with which to model very large and non-symmetrical chemical systems such as proteins and polymers. Molecular mechanics is a purely empirical method that neglects explicit treatment of electrons, relying instead on the laws of classical physics to predict the chemical properties of molecules . As a result, MM calculations cannot deal with problems such as bond breakage or formation, where electronic or quantum effects dominate. Furthermore, MM models are wholly system-dependent. MM energy predictions tend to be meaningless as absolute quantities, as the zero or reference value depends on the number and types of atoms and their connectivity, and so they are generally useful only for comparative studies. A force field is an empirical approximation for expressing structure-energy relationships in molecules and

Molecular mechanics have been shown to produce more realistic geometry values for the majority of organic molecules, owing to the fact that they are highly parameterised. Parameterisation of structures should be performed with care and non-"standard"

**2.1. Molecular mechanics, molecular dynamics and docking** 

is usually a compromise between speed and accuracy.

Tóth et al., 2005].

hydrophobic effect).

To apply computational methods in drug design, it is always necessary to remember that to be effective, a designed drug must discriminate successfully between the macromolecular target and alternative structures present in the organism. The last few years have witnessed the emergence of different computational tools aimed at understanding and modelling this process at the molecular level. Although still rudimentary, these methods are shaping a coherent approach to help in the design of molecules with high affinity and specificity, both in lead discovery and in lead optimisation. Moreover, current information on the 3D structure of proteins and their functions provide a possibility to understand the relevant molecular interactions between a ligand and a target macromolecule. As a consequence, a comprehensive study of drug structure–activity relationships can help identify a 3D pharmacophore model as an aid for rational drug design, as a pharmacophore model can be defined as 'an ensemble of steric and electronic features that is necessary to ensure the optimal supramolecular interactions with a specific biological target and to trigger (or block) its biological response', and a pharmacophore model can be established either in a ligandbased manner, by superposing a set of active molecules and extracting common chemical features that are essential for their bioactivity, or in a structure-based manner, by probing possible interaction points between the macromolecular target and ligands.

Molecular recognition (MR) is a general term designating non-covalent interactions between two or more compounds belonging to host-guest, enzyme-inhibitor and/or drug-receptor complexes. A rigorous approach to an MR study should involve the adoption of a computational method independent from the chemical intuition of the researcher. Drug design purposes prompt another challenging feature of such an ideal computational method, the ability to make sufficiently accurate thermodynamic predictions about the recognition process.
