Domain Interactions as Potential Drug Targets

Protein-protein interactions preside over the physiological and pathological processes in living organisms. Therefore, the ability to intervene in a controlled manner in the protein interaction network, altering its wiring by specifically favoring certain interactions and blocking others, is crucial for molding the cell response into a desired phenotype. Examples of the efficacy of such strategy is witnessed by the modus operandi of some known pathogens, such as the Epstein-Barr virus (EBV) (Longnecker and Miller 1996) and the enteropathogen E. coli (Gruenheid et al. 2001): specific viral and bacterial proteins possess recognition motifs that are able to recruit host proteins, thus reorganizing the wiring of the protein network to better fit the pathogen needs (Pawson et al. 2001). In these cases, the perturbation of the cellular interaction network leads to disease onset.

The outstanding importance of a tight regulation of protein interactions, governed by the underlying domain interactions, for proper cell functioning has led scientists to regard them as prominent targets for drug design. Research in this direction has been carried on the following two complementary approaches: (1) designing molecules to induce protein-protein interactions (dimerizers); (2) designing molecules to inhibit protein-protein interactions (inhibitors).

Since many signaling pathways are initiated by a dimerization event, pathway activation may be stimulated artificially by small molecules endowed with anchors having affinity for the protein partners that are to be brought into proximity (Archakov et al. 2003). The anchors are separated by a linker region of variable length. To avoid undesirable side effects and increase dimerizer specificity, a mutant domain may be inserted in the construct expressing the wild-type target proteins so as to properly tune their affinity to the anchor part of the interfering molecule. Over the last decade, several dimerizers have been proposed and tested in the laboratory (Austin et al. 1994; Clemons 1999; Belshaw et al. 1996; Kopytek et al. 2000). Theoretically, dimerizers could be used in cancer therapy to induce apoptosis in abnormally proliferating cells (Amara et al. 1997; MacCorkle et al. 1998).

When the challenge of inhibiting the formation of a protein complex is faced, different strategies must be adopted depending on the type of protein contact surface we are dealing with. As we have seen in paragraph 0, the domain-domain interfaces of permanent complexes have different physical and chemical characteristics than those of temporary complexes. Namely, the shaping of contact surfaces in permanent oligomers involves the burial of large areas of the protein surface within the complex core, resulting in a tightly packed structure held together by strong non-covalent interactions. This process can be thought of as a continuation of protein folding (Archakov et al. 2003). The high-energy bonds involved in the interaction and the high level of complementarity between the domain surfaces make it difficult to design small molecules that would effectively block the formation of the macromo-lecular complex by competitive binding. However, some degree of success has been reached thanks to a mechanism called "interfacial inhibition," whose main idea is to trap macromolecules into dead-end complexes while they are undergoing conformational changes by letting artificial or natural compounds bind to "hot spots" that are present during folding transition states only. For instance, it has been shown that two natural inhibitors, brefeldin A (BFA) and camptothecin (CPT), have the power to freeze the targeted complexes (respectively, an Arf-GDP-Sec7 complex and a TOP1-DNA complex) in a permanently inactive conformation, preventing them from completing their biological function (Pommier and Cherfils 2005).

On the contrary, transient interactions nucleating the assembly of temporary protein complexes resemble receptor-ligand interactions. Many of these interactions are mediated by protein interaction modules binding to short peptides exposed on the surface of the target proteins. The contact area between the recognition pocket of the interaction module and the target peptide is usually small, allowing for competitive disruption of the complex. Using site-directed mutagenesis, it has been possible to determine the amino acid residues important for binding, and thus peptides resembling the dimer interface have been designed. The binding ability of the construct can be subsequently optimized by combinatorial chemistry techniques in order to individuate the amino acid composition that would confer to the synthetic construct a binding affinity that is possibly higher than that of the physiological partner. Peptidomimetics and small organic molecules have also been tested as inhibitors with good success rates (Zutshi et al. 1998; Jiang et al. 2006; Arkin and Wells 2004; Turkson et al. 2004).

Due to their pervasiveness and fundamental role in biological processes, protein interactions certainly hold enormous potential as drug targets. With respect to antibiotic design, targeting protein-protein interfaces rather than enzyme active sites has evident advantages: protein-protein interfaces have structural peculiarities that facilitate discrimination between the host proteins and the pathogen ones, whereas enzyme active sites show little difference. As a consequence, a potential drug directed against a pathogen enzyme active site may end up blocking the action of the cell's own enzymes (Singh et al. 2001). Furthermore, whereas pathogens can acquire antibiotic resistance by mutating a few residues in an enzyme active site, while maintaining the catalytic ones unaltered, in order to neutralize the action of a protein interaction inhibitor the pathogen would have to mutate simultaneously a few amino acid positions in both interacting surfaces to preclude drug binding and, at the same time, to preserve the ability of the partner proteins to interact. This is, of course, statistically more improbable.

0 0

Post a comment