ИСТИНА

Structures of complexes between proteins and nucleic acids (NA) are widely used indesign of molecular biology experiments. Nevertheless, data from a single structurecould not always provide information concerning the most important and conservedfeatures of protein-NA interactions. To perform deep and comparative analysis ofprotein-NA complexes, the Nucleic acid—Protein Interaction DataBase (NPIDB)can be used. NPIDB currently contains 5440 structures of DNA-protein and RNA-protein complexes supported with search and navigation tools. For any structurefrom NPIDB it is possible to download PDB-structure; ĕnd hydrophobic clusters,SocBiN Bioinformatics 201641 hydrogen bonds and potential water bridges in protein-NA complex; view sequences,view 3D structure in Jmol, link to Pfam and SCOP. Original structural classiĕcationof DNA-binding modes of protein domains is implemented in NPIDP. For each an-notated protein, family superimposition of structures, sequence alignment with sec-ondary structure, conserved interface water clusters and interaction class are available.Based on this data comparative analysis of protein-DNA complexes can be done. Sug-gested classiĕcation deals with the structural protein domains but not with the wholeDNA-recognized proteins. It allows predicting potential protein-DNA contacts forstructures crystallized without DNA or for proteins which combine several DNA-recognizing domains. Here, using a family of TATA-box binding proteins (TBPs)and LAGLIDADG_1 proteins as an example, we present workĘow to analyze pro-tein domains and whole protein complexes with nucleic acids using services and datafrom NPIDB. e workĘow included: (1) creation of structural superimposition andsequence alignment of TBPs and LAGLIDADG-DNA complexes which were gener-ated by PDBeFold service or were obtained from NPIDB; (2) determination of con-served hydrogen bonds and hydrophobic interactions based on structural superim-position, sequence alignment and data of hydrogen bonds and hydrophobic interac-tions for each TBP and LAGLIDADG-DNA complex; (3) calculation of conservedwater molecules clusters with wLake program (for the whole protein) or obtaining ofthese clusters from NPIDB (for domains); (4) Creation of interaction scheme for ana-lyzed protein-DNA complexes based on the above. Nine amino acid residues makingconserved hydrogen bonds, 13 residues participating in formation of two conservedhydrophobic clusters at DNA-protein interface, and four conserved water-mediatedcontacts were found for TBP proteins. Partial symmetry of conserved contacts re-Ęects quasi-symmetry of TATA-box protein binding structure. In contrast with TBPs,structures of LAGLIDADG proteins do not contain conserved amino acids, but theirconserved amino acid positions (14 in total) interact with DNA. While hydrophobicclusters mediate protein interactions with DNA backbone, direct and water-mediatedhydrogen bonds form conserved contacts with DNA backbone. Taking into accountconserved contacting positions instead of amino acid residues can be useful for ho-mologous proteins with low sequence similarity.