Pharmacophoric features

Methods that deduce a pharmacophore, an arrangement in 3D space of features that contribute or detract from binding and look for its presence in the database that is searched. This method places emphasis on features like hydrogen bond donors, hydrogen bond acceptors, acidic or basic units and hydrophobic fragments and opens the possibility of identifying unexpected scaffolds with required features (pharmacophore-based VS or PHBVS). 

Figure 6.3 Common-feature pharmacophores ofala adrenergic receptor antagonists [16]. Onto each pharmacophore the reference has been mapped, (a) Class I pharmacophore model aligned to prazosin, (b) class II pharmacophore model aligned to compound 10.

For sufficiently similar models, this algorithm is extremely useful for deriving common-feature pharmacophores. Its use, however, is restricted to the nearly identical binding pockets conformational differences in the two compared protein pockets will lead to useless results. 

To test this hypothesis, we investigated a new series of phenylpiperazines, bearing a pyrimido[5,4-d]indolo group as the terminal heterocyclic moiety like compound 5, showing preferential affinity for the ald-AR subtype. A selection of 16 compounds, with affinity values spanning over 3.5 orders of magnitude, was used to generate a set of 10 five-feature pharmacophore hypotheses [18] that we compared with the previously described five-feature pharmacophoric... 

Fig. 12.3 Comparison between the first six-feature pharmaco-phoric model for a1(j-AR antagonists and the improved five-feature pharmacophoric model for the same arAR subtype antagonists.

All these facts lead to the conclusion that our new five-feature pharmacophore model should be a good abstract representation of the most important structural elements that a compound should possess for high a1(j adrenoceptor affinity. Similarity between the HBA and PI features of the pharmacophore model with parts of the theoretical receptor could be also considered as an additional validation of our model. 

In a paper published in 2000 by Norinder [30], Catalyst was used for the first time to build a common feature pharmacophore hypothesis for HIV-1 protease inhibitors, which was then refined using in-house software (HypoOpt), after having added to it some hundreds of excluded volume spheres. These were actually derived from the X-ray structure of an inhibitor complexed to the enzyme. The aim of the approach was to obtain a computational model with some improved predictive power with respect to the corresponding hypothesis derived without receptor information. 

Steindl, T., Langer, T. Influenza virus neuramidase inhibitors generation and comparison of structure -based and common feature pharmacophore hypotheses and their application in virtual screening. J. Chem. Inf. Comput. Sci. 2004, 44, 1849-1856. 

The two common-feature pharmacophores are depicted in Fig. 13.3a and b showing the mapping on to a reference molecule representative for both classes of high affinity a1A antagonists [11]. The models describe the key chemical features required for binding of structurally diverse ligands to this adrenergic re-... 

Objective Find common feature configurations amongst a set of active molecules. Algorithm HipHop uses a pruned exhaustive search method. Starting with simple two-feature pharmacophores, the program tries to add one extra common feature at a time until no larger common pharmacophore configuration exists [10]. Combinations that cannot be completed to reach a minimum number of features are not further explored. 

For database searching, pharmacophores are best defined by all possible distances between chosen groups or features (pharmacophore points). Therefore, as illustrated in Figure 1.13, they are best represented as a molecular graph (similar to reduced graphs). In this case, different from conventional graphs, however, nodes correspond to points (or centroids) and edges to inter-point distances, rather than bonds. 

Schuster, D., Maurer, E.M., Laggner, C., Nashev, L.G., Wilckens, T, Langer, T, and Odermatt, A. (2006) The discovery of new 1 Ibeta-hydroxysteroid dehydrogenase type 1 inhibitors by common feature pharmacophore modeling and virtual screening. Journal of Medicinal Chemistry, 49, 3454-3466. 

---