FXa is a vitamin K

FXa is a vitamin K-dependent serine protease consisting of two chains linked by a disulfide bridge. The heavy chain contains 303 nm to lb australia and the light chain has 139 amino acids. The catalytic triad comprised of Ser195, His57, and Asp102  exhibits the trypsin-like β-barrel structure [12]. The active site of FXa contains four pockets (Fig. 1) [[13], [14], [15]], but there are controversies concerning the exact assignment of specific residues to certain pockets. S1 pocket is considered to be the most crucial for ligand binding [16] and represents the negatively charged deep site consisting of Ser195, Tyr228, and Asp189. S3 pocket is small and formed by Gln192. Some authors attribute Gln192 to S1 pocket and don't distinguish S3 pocket [17,18]. S4 pocket consists of Phe174, Tyr99, and Trp215 [14,16,18] – residues which carry aromatic side chains and thereby make this pocket capable of having an affinity to hydrophobic moieties. Furthermore, this aromatic pocket is responsible for selective retention of inhibitors [14,18]. S2 pocket of FXa can be distinguished by CoMFA analysis [13], but, in fact, access to this pocket is blocked by Tyr99 [18]. Another important element of FXa active site is a β-strand region defined by the 214–218 backbone that contains Gly216 and Gly218 – residues frequently implicated in formation of hydrogen bonds with a ligand [19]. Some authors consider Gly216 to be related to S1 pocket [13,16]. Several FXa inhibitors with moderate activity were previously identified by our research group [20]. The approach which was used in this identification represented docking-based virtual screening of two large chemical libraries. The found inhibitors had different scaffolds and the best found inhibitor, 17f, was a derivative of pyrroloqinoline which showed IC50 = 0.7 μM (Fig. 2). Results of this screen were exploited for generating the virtual focused library of compounds which contained scaffolds similar to ones the identified inhibitors have. In silico analysis of the library was supposed to result in discovery of more potent inhibitors than those found previously. In the present study, we continued the previously published work aimed at designing inhibitors for FXa. Here, we have applied more thorough and more accurate approach for in silico prediction of activity than the previous one was. The main objective was to find inhibitors with the considerable activity against FXa and to explore structural analogs of previously found FXa inhibitors. In the first part, the screening of the focused library consisting of 244 compounds, which were designed as modified analogs of FXa inhibitors identified at the previous stage [20], was conducted by using the methodology combining docking and semiempirical quantum chemical postprocessing. It identified 20 candidates, which, then, were synthesized. Experimental testing based on an amidolytic assay has indicated that 7 of them were active against FXa at maximal concentrations, allowed by solubility, and one compound had the IC50 = 1,9 μM. This hit compound contained 1,2,3,4-tetrahydro-2,2,4,7-tetramethylquinoline as a scaffold. In the next iteration, we therefore studied additional derivatives of 1,2,3,4-tetrahydro-2,2,4,7-tetramethylquinoline and selected 16 additional candidates for inhibiting FXa. Of these, 3 compounds showed noticeable activity at 5 μM. Finally, we performed the screen of the Voronezh State University database by means of the same computational approach and selected only those compounds which had 1,2,3,4-tetrahydroquinoline scaffold. This screen resulted in selection of 4 candidates, among which two additional FXa inhibitors were identified by in vitro testing. One of the found inhibitors showed IC50 = 3 μM. All of the six best substances were specific to FXa over thrombin, three of them showed specificity over trypsin and one substance were also specific over factors IXa and XIa.