The tetracycline-binding site of the TetR Selumetinib protein has been defined and the structure determined by crystallography. We found that the TetR protein shares similar characteristics with the E protein in the binding sites for the tetracycline derivatives. First, there is an appropriate volume in the binding sites. The volumes of the binding sites of various TetR crystals range from 359 A ? 3 to 495 A ? 3 whereas the BOG binding site on the E protein is 481 A ? 3, according to the tool program, QSiteFinder. Therefore, there is proper space for the tetracycline derivatives to fit into the BOG binding site. Second, there are hydrophobic surfaces in the pockets of both binding sites. Third, according to the results of a cross-docking test performed for TetR and the tetracycline derivatives, the binding sites of the DV E protein and TetR permit the binding of the tetracycline derivatives. In addition, the hydrogen bonds formed between the tetracycline derivatives and the DV E protein are similar to those between TetR and the tetracycline-derived ligands. Therefore, tetracycline derivatives should reasonably bind the BOG pocket of the DV E protein. On the other hand, only two of the derivatives are inhibitory; therefore, the atomic details of the functional ICI 182780 groups and the tetracyclic core must confer the inhibitory activity. Hence, we have analyzed the docked conformations and hydrogen bonding of the derivatives to assess the interaction between those compounds and the E protein. There are distinct differences between the effective and ineffective compounds; the effective compounds have their tetracyclic cores positioned inside the pocket while their side chains form hydrogen bonds with the residues located on the opposite sides of the wall around the pocket and are capable of creating steric hindrance to the conformational alteration of the E protein. In contrast, the ineffective compounds form hydrogen bonds only with one side of the wall and their cores lean away from the pocket. Next, on anatomiclevel, the predictedpositions of thetetracycline derivatives with the E protein are shown in Figures 6 and 7. The fused tetracyclic rings of rolitetracycline and doxytetracycline bind along the D9o strand and occupy the D9c space of the E protein. The residues 48�C52 are in the D segments. These compounds both interact mainly with Thr48, Glu49, Ala50, Gln200, and Gln271 through hydrogen bonds. Such a hydrogen-bonding network provides strong attraction forces to stabilize the binding of rolitetracycline and doxytetracycline to the D9o strand and the kl b-hairpin. In contrast, although these compounds have the same tetracyclic core structures, neither tetracycline nor oxytetracycline is inhibitory. Both compounds form hydrogen-bonding networks with Thr48, Gln200, Gln271, Phe279, and Thr280; therefore, their tetracyclic rings are docked toward one side of the binding site and contact the surrounding hydrophobicresidues via van der Waals interactions, which are very different from those of rolitetracycline and doxytetracycline. During the process of E protein-host membrane fusion, the E protein structure is dramatically re-configured to allow the fusion peptide to properly interact with the host membrane. This event is marked by the rearrangement of the kl b-hairpin and the D9o segment in the BOG binding site. The docked positions of the inhibitors suggest that they occupy the D9c and kl b-hairpin spaces in the post-fusion state and form a stable hydrogen-bonding network. Therefore, these compounds block the rearrangement of the b-hairpin and D9o strand, and thereby block the rearrangement of domains II and I of the E protein during membrane fusion. Residues 48-52 are not only important to inhibitor binding but may also directly affect flavivirus membrane fusion. This hypothesis is consistent with previous reports that Gln52 may affect the pH threshold of fusion in flaviviruses. Our study has presented a cost-effective and time-saving screening process that is based on limited structural information.