Evolutionary relationships among proteins probed by an iterative neighborhood cluster analysis (INCA). Alignment of bacteriorhodopsins with the yeast sequence YRO2.

PURPOSE

Searching the existing databases for homologous sequences is essential to understanding a protein's structure and function. For a query sequence, its nearest neighbors can be identified by BLAST (basic local alignment search tool). However, a single query sequence is sufficient to define the entire neighborhood of related sequences, and multiple BLAST queries are needed. We describe here a program which permits automated and iterative BLAST analysis of an entire neighborhood of sequences and apply this to search for homologs of the bacteriorhodopsins outside the archaea phylum.

METHODS

We have developed a Java program, 'Iterative Neighborhood Cluster Analysis' (INCA), which performs iterative BLAST searches, beginning with a single starter sequence, and proceeding with any other sequence achieving a predefined minimum alignment score. This results in a cluster of sequences where each sequence is related to at least one other sequence by the cutoff score, additional lists of more distantly related sequences for each member of cluster.

RESULTS

Bacteriorhodopsins had not been previously aligned with any other protein family with scores indicative of probable homology. Using INCA, we identified a probable homolog in yeast, YRO2_YEAST, also containing seven putative transmembrane domains. A finding of probable homology was supported by additional alignment strategies.

CONCLUSIONS

INCA is a useful tool to assess complete protein neighborhoods. With an increasing database, INCA can serve to detect the emergence of evolutionary links between even the most distantly related protein families. Identifying a homolog of the bacteriorhodopsins in yeast illustrates this approach but at the same time highlights the vast evolutionary distances between polytopic membrane proteins, such as the bacteriorhodopsins.