3D-QSAR modeling and pharmacophore study of serotonin 5HT-₂A receptors antagonists

Abstract

Creation of a statistically reliable 3D-QSAR (Quantitative Structure Activity Relationships) model enables definition of the structure of the pharmacophore for antagonists of serotonin 5-HT2A receptors and selection of the key molecular determinants for the design of new antagonists. A 3D-QSAR study was applied on a series of 50 antagonists of 5-HT2A receptors. The data set is divided into the training set, composed of 33 compounds, and test set containing 14 molecules. The training set was used to build the 3D-QSAR model, while test set was used for model validation. This 3D-QSAR study was performed by use of the Pentacle 1.07 programs. Obtained statistical and validation parameters for training set (R²= 0.96; Q²=0.75, RMSEE= 0.211); and for test set (R²pred=0.51; RMSEP= 0.558); indicate on reliability and good predictive potential of the 3D-QSAR model. Based on the most influential variables of the selected 3D-QSAR model, the molecular determinants for the antagonistic effect on the 5-HT2A receptors were selected: hydrogen bond acceptor and hydrogen bond donor at a distance of 10.4A-10.8A (v495: O-N1); carbonyl oxygen and a steric hot spot at a distance of 14.8A-15.2A (v640: N1-TIP); hydrophobic domain and hydrogen bond acceptor at a distance of 3.2A-3.6A (v276: DRY-O); two steric hot spots at a distance of 18.8A-19.2A (v248: TIP-TIP) .

Formiranjem statistički pouzdanog 3D-QSAR (Quantitative Structure Activity Relationships) modela omogućeno je definisanje strukture farmakofore antagonista serotoninskih 5-HT₂A receptora i selekcija ključnih molekulskih determinanti za dizajn novih antagonista serotoninskih 5-HT₂A receptora. 3D-QSAR studija je primenjena na seriji od 50 antagonista serotoninskih 5-HT₂A receptora koji su podeljeni na trening set od 33 molekula i test set od 14 molekula. Trening set je korišćen za formiranje 3D-QSAR modela, dok je test set primenjen za validaciju modela. Za ovu 3D-QSAR studiju je upotrebljen Pentacle 1.07 program. Izračunati validacioni i statistički parametri 3D-QSAR modela (R²=0,96; Q²=0,75; RMSEE= 0,211), kao i parametri eksterne validacije na test setu (R²pred =0,51; RMSEP= 0,558), ukazuju na pouzdanost i dobru moć predviđanja izabranog 3D-QSAR modela. Na osnovu najuticajnijih varijabli izabranog 3D-QSAR modela definisana je struktura farmakofora za antagonističko dejstvo na serotoninskim 5-HT₂A receptorima: donor i akceptor vodonične veze na rastojanju od 10,4A-10,8A (v495: O-N1); karbonilni kiseonik i sterni centar na rastojanju od 14,8A-15,2A (v640: N1-TIP); hidrofobni centar i donor vodonične veze na rastojanju od 3,2A- 3,6A (v276: DRY-O); dva sterna centra na rastojanju od 18,8A-19,2A (v248: TIP-TIP).