RAPISARDA Viviana Andrea
congresos y reuniones científicas
Strategies to recognize protein subfamilies: revision of NADH dehydrogenase-2 classification within the pyridine nucleotide-disulphide reductases
ÁVILA, C. L.; RAPISARDA, V. A.; FARIAS, R. N.; DE LAS RIVAS, J.; CHEHÍN, R.
Congreso; Segundas Jornadas Iberoamericanas de Bioinformática; 2006
Red Iberoamericana de Bioinformática
The pyridine nucleotide-disulphide reductase (PNDR) is a large and heterogeneous family divided into two classes reflecting the divergent evolution of its characteristic disulfide redox-active site. However, not all PNDR members clearly fit in this classification, suggesting the necessity of a reclusterization of this family. Fingerprints are flexible and powerful tools useful in protein family classification. A novel framework to obtain a fingerprint guided clusterization of protein sequences was designed and further applied on the PNDR family. Obtained clusters were compared with those derived from phylogenetic analysis showing similar results. Two large groups, which correspond to the previously described PNDR Class I and Class 11, were distinguished by both methods. Still, two other separate clusters. which were previously classified as Class I on most databases, were also identified: NADH dehydrogenase-2 and peroxide reductases. Since these proteins have acquired a different redox active site. It appears reasonable to reclassify them as separate classes. Further analysis combining classical bioinformatics tools and expert knowledge derived from biochemical data were also performed on the NADH dehydrogenase-2 protein family. This analysis identified a signature formed by a combination of specific motifs. which may be related to the observed copper reductase activity. The present strategy proved to be efficient or the classification of protein families, reflecting the evolutionary relationship among their members. The analysis performed on PNDR family led to a revision of its clusterization. In addition, the described fingerprints could also be useful for retrieving sequences from databases.