Introduction

The idea behind biostructure-based drug design is to utilize the information on shape and properties of the binding site of a target molecule (e.g., enzyme or receptor) to design compounds, which possess complementary properties. Thus, biostructure-based drug design requires methods for determination of the three-dimensional (3D) structure of the target molecules as well as knowledge of which molecular interactions are important to obtain the desired binding characteristics.

Examples of ligands (drug molecules) binding to proteins are shown in Figure 2.1. The two ligands have been selected to illustrate different types of molecular interactions between the ligand and the target protein.

The 3D structure of a target protein can be determined experimentally by methods like x-ray crystallography and NMR spectroscopy, or predicted by computational methods like homology modeling (comparative model building). Of 50,000 experimentally determined protein structures 43,000 have been determined by x-ray crystallography (Protein Data Bank, May 2008). An x-ray crystallographic structure determination requires protein crystals, and irradiation with a high-energy x-ray source generates a diffraction pattern by the scattering of x-rays from organized molecules in a continuous arrangement in the crystal. Based on the diffraction pattern, an electron density map of the protein can be derived and subsequently a molecular model reflecting the electron density, the 3D structure, can be determined. Presently, the data collection, data processing, model building, and refinement are highly automated and computerized processes. The present limiting factor for determining the 3D structure of a protein is to get sufficient amounts of pure and stable protein and proper diffracting crystals.

It is important to consider the quality of an x-ray structure before using it for biostructure-based drug design. The resolution is a measure of how detailed the electron density map is and thereby how accurately the positions of the individual atoms can be determined (Figure 2.2A). Structures based on electron densities at 1.2 A resolution are normally referred to as atomic-resolution structures and, e.g., hydrogen-bonding networks can unambiguously be identified. Generally, a resolution of ca. 2 A

FIGURE 2.1 Example of a nonpolar (A) and a polar (B) binding site. In (A), the ligand, which is an analog (R = Br) of the COX-2 inhibitor celecoxid (R = CH3), binds to cyclooxygenase-2 in a pocket primarily formed by nonpolar amino acid residues (pdb-code 1CX2). In (B), the ligand, which corresponds to the active part of the anti-influenza drug oseltamivir (Tamiflu), binds to an influenza virus neuraminidase in a pocket formed by polar residues (pdb-code 2QWK). Green arrows indicate hydrogen bonds. Green and red circles represent nonpolar and polar residues, respectively. The dotted lines illustrate the shape of the binding sites.

FIGURE 2.1 Example of a nonpolar (A) and a polar (B) binding site. In (A), the ligand, which is an analog (R = Br) of the COX-2 inhibitor celecoxid (R = CH3), binds to cyclooxygenase-2 in a pocket primarily formed by nonpolar amino acid residues (pdb-code 1CX2). In (B), the ligand, which corresponds to the active part of the anti-influenza drug oseltamivir (Tamiflu), binds to an influenza virus neuraminidase in a pocket formed by polar residues (pdb-code 2QWK). Green arrows indicate hydrogen bonds. Green and red circles represent nonpolar and polar residues, respectively. The dotted lines illustrate the shape of the binding sites.

FIGURE 2.2 (A) Electron density of a phenylalanine side chain at 1.1, 2.0, and 2.8 A resolution, respectively. (B) Ramachandran plot of a protein. Two residues (Lys76 and Leu178) adopt unfavorable backbone conformations. The majority of the residues are located in the dark gray regions corresponding to favorable backbone conformations.

FIGURE 2.2 (A) Electron density of a phenylalanine side chain at 1.1, 2.0, and 2.8 A resolution, respectively. (B) Ramachandran plot of a protein. Two residues (Lys76 and Leu178) adopt unfavorable backbone conformations. The majority of the residues are located in the dark gray regions corresponding to favorable backbone conformations.

ooooooooo ooooooooo ooooooooo ooooooooo ooooooooo ooooooooo ooooooooo ooooooooo ooo^""roo ooo ; Joo ooo ^b o o ooooooooo ooooooooo ooooooooo ooooooooo ooooooooo ooooooooo o a o a

o ooooooooo ooooooooo o a o a

o ooooooooo ooooooooo ooooooooo ooooooooo ooo^""roo ooo ; Joo ooo ^b o o ooooooooo ooooooooo poo 000

ooooooooo ooooooooo

FIGURE 2.3 Simplified binding process. (A) Desolvation of protein and ligand (large unfavorable energy change). (B) The binding process (small favorable energy change). (C) Solvation of the protein-ligand complex (large favorable energy change).

provides an accurate structure, and accordingly such a structure is suitable for biostructure-based drug design. Structures based on 3 A resolution electron densities should be used with caution.

The stereochemical quality of a protein structure should be carefully evaluated prior to using it for biostructure-based drug design. Planarity of peptide bonds, bond lengths, bond angles, and torsion angles should not deviate from average values. The protein backbone geometry is usually validated by a Ramachandran plot, a plot of the two variable torsion angles (phi and psi) (Figure 2.2B). Special attention should be paid to the part of the protein involved in direct interactions with ligands, e.g., flexible loops and binding site residues.

Binding of a drug molecule to a protein is a complicated process which is often considered as composed of a number of discrete steps in order to simplify the understanding of the process: (1) the protein may contain a cavity where the drug may bind, but prior to binding the water molecules occupying the cavity has to be removed. Similarly, the drug molecule may be surrounded by water molecules that have to be removed before binding. These two desolvation processes are associated with an unfavorable energy change (Figure 2.3A). (2) The next step is the actual binding between the drug molecule and the protein. The drug molecule may change conformation in order to fit into the binding site of the protein, and this conformational change requires energy. The protein may also change conformation, but this is usually ignored. The binding process requires that the drug molecule and protein are complementary not only with respect to shape, but also with respect to molecular properties. Positive parts of the drug molecule will bind to negative parts of the binding cavity and vice versa, and hydrogen-bond donors will bind to hydrogen-bond acceptors and vice versa (Figure 2.3B). (3) Finally, the complex between the drug molecule and the protein has to be surrounded by water molecules (solvated), and this process is associated with a favorable energy change (Figure 2.3C). Thus, the actual binding energy is a relatively small difference (typically on 5-20 kcal/mol) between the processes representing large destabilizing vs. stabilizing forces (for further details see Chapter 1).

Was this article helpful?

0 0

Post a comment