From Protein Structure to Function with Bioinformatics.pdf

260 R.A. LaskowskiFig. 10.4 Schematic diagram of the sequence- and structure-based methods used in ProFunc. Thesequence-scans in the left-hand column include searches against the protein sequences in the PDBand uniprot databases. The interproscan and Superfamily searches return any sequence motifs fromtheir databases that are present in the query protein’s sequence. For each uniprot BLAST hit thegene neighbours search locates the gene in its genome, if available, and identifies all neighbouringgenes. The first of the structure-based searches in the middle column uses SSM to identify structureswith the most similar overall fold to that of the query protein. Surface clefts are computed andcan be visualized coloured by residue type or residue conservation. Two types of structural motifsare identified: helix-turn-helix (HTH) motifs, characteristic of many DNA-binding proteins, andnests, which are often found at functionally important locations. Finally, in the right-hand column,are the various template methods that find local 3D matches to known protein structuresture elements (SSEs) of the target structure against those of the structures in thedatabase. Any strong matches are superposed and an r.m.s.d. for equivalent Cαs iscalculated together with a z-score measure of significance and SSM’s own significancemeasure, called the Q-score. In ProFunc the top ten hits, ordered by Q-score,are shown and any, or all, can be viewed superposed on the target structure usingthe molecular graphics viewer RasMol (Sayle and Milner-White 1995).The top fold match to our example structure, 2fck, is shown in Fig. 10.5. Thematch is to PDB entry 1s7f, a RimL N(α)-acetyltransferase from Salmonella typhimurium(Vetting et al. 2005). This protein is responsible for converting the prokaryoticribosomal protein from L12 to L7 by acetylation of its N-terminalamino group. The protein forms a homodimer and the dimer interface creates a