Fig. 14A, B Equilibrium between the two conformations of IQA present in solution. As evidenced by nuclear magnetic resonance (NMR) analysis, the species with the internal hydrogen bond is predominant in solution (species A on the left)

three hydrogen bonds, one with the amine function of Lys68, one with the backbone nitrogen of Asp175, and one with a water molecule in turn connected to the backbone amide of Trp176 (Fig. 13). This water molecule is the same found in all CK2 structures solved to date, with the exception of the emodin complex, as already mentioned.

For both conformations, the major contribution to the binding comes from the hydrophobic interaction with non-polar residues in the binding site, i.e., Val45, Val53, Ile66, Lys68, Val95, Phe113, Val116, Met163, and Ile174. In this respect, it is remarkable that the total buried surface upon inhibitor binding is quite large, about 730 A2. The importance of the hydro-phobic interactions in the binding of IQA is supported by the considerable decrease of the inhibitory efficiency observed in the cases of Val66Ala and Ile174Ala mutants. In fact, mutation of Ser51 with a glycine does not significantly affect the Ki value. This is in agreement with the observation that the two different orientations of IQA bind with similar affinity, since they share the hydrophobic contacts, while they differ in the interaction with Ser51.

The best inhibitor of CK2 discovered so far, IQA, albeit more potent and selective than those previously available, is not yet totally specific for the enzyme. With the increasing number of three-dimensional structures of in-hibitors/CK2 complexes, it should be possible to improve the selectivity of inhibitors using a structure-based drug design method. Anyhow, this task is not straightforward, since we have seen from the previous examples that compounds extremely similar to one another, like MNA, MNX, DAA, and emodin, bind to the active site of the enzyme in a significantly different way. In practice, a limited modification of the structure of the inhibitor, for example the addition of a substituent with the aim of increasing the number of H-bond interactions (like the addition of an OH or NH2 group), often changes the electronic properties of the inhibitor, modifying its binding mode. In the case of CK2 inhibitors, this is probably worsened by the rela tively limited complexity of the small molecules used. A more extended surface of interaction between the inhibitor and the enzyme could help not only to decrease the Ki, but also to stabilize its binding mode.

Another point about computer-aided drug design has to be stressed. In the in silico procedure for the identification and characterization of the indoloquinazolinone derivatives, such as CK2 inhibitors, a model of IQA bound to the enzyme has been proposed (Vangrevelinghe et al. 2003). In this model, the acetate group of IQA faces the hinge region and interacts with ar-ginine-43, with a orientation opposite to that found in the experimental crystal structure. In this case, it comes out that while virtual screening protocols could usefully select some interesting compounds among a large database, the docking procedures have been unable to correctly position the inhibitor in the enzyme active site. For the correct detailed description of the interactions between IQA and CK2, the determination of the crystal structure of the complex was an unavoidable step.

0 0

Post a comment