Dissertation is devoted to the library creation of potential ligands for SH2 domains and their biological verification. All available SH2 domains and their complexes with ligands were analyzed, 56 structures with unique amino acid sequences were identified and divided into 6 groups and 8 representative binding pockets. Comparable domain structures are characterized by high/average levels of conservation (group 1 - 92/91%, group 2 - 81/87%, group 3 - 56/62%, group 4 - 37/51%, group 5 - 50/68% and group 6 - 51/61%) and not significant Rmsd difference (0.8 A, 1.16 A, 1.6 A, 1.9 A, 2 A and 1.4 A). It is established that the binding site is more conservative (94/93%, 89/96%, 59/67%, 49/60%, 53/70% and 53/59%, respectively) and comprises 20-24 key amino acids: FLVRESETT (pTyr-binding part), ?B-sheet; KHYKIR (central part), ?D-sheet; ITSR and ADGLC (hydrophobic part of binding pocket), ?G-fold and ?B-helix. Four key points of ligand binding pocket of SH2 domains were found: (1) H-bond acceptor (pTyr-binding part), (2) aromatic/hydrophobic part, (3) H-bond donor/acceptor, (4) hydrophobic part. Also, four "SH2-domain-ligand" docking models were received: almost in all cases ligand tightly fills in binding site. In general, three library of potential ligands to SH2 domains were obtained, which are containing 1015, 3018 and 10463 structures, respectively. By using active compounds to SH2 domains 9 pharmacophore models were built. Their virtual screening showed good quality parameters, including the area under the ROC curve (AUC) varies between 75% and 100%. Thus, 1816 potentially active substances were selected. GTM model was created for predicting the activity of potential ligands too. Several different chemotypes, which can mimic phosphorylated tyrosine, were identified and as a result 1607 potentially active compounds were selected. In biological test the inhibitory effects of potential ligands of SH2 domains on proliferation of tumor cells and kinase activity of Btk and Syk kinases were evaluated.
